# Glossary *This glossary brings together the essential terms and concepts that define our R\&D practice. With 100 entries, it serves as both a reference guide and a learning resource to help you understand what these terms mean, explain them to others, and see how they connect forming a body of knowledge. This is not an exhaustive list of terms; it focuses on the primary concepts contained in this guide and those that have emerged as most essential to our practice. The footnotes provide additional resources if you want to explore any concept in greater depth, and may help you in case you need to adapt these definitions to your local context.* ### ​​Accelerator Lab Network The UNDP Accelerator Lab Network is a global learning network working within the United Nations Development Programme (UNDP) to accelerate progress on sustainable development. Started in 2019 to reimagine sustainable development, it has evolved into an open, globally distributed R\&D capability for the Sustainable Development Goals. The Labs combine local knowledge with new data, digital tools, and experimental methods to address complex challenges in their regions. Currently, there are 89 Labs in 113 countries. They help UNDP and its partners find and share effective solutions while also learning how to operate effectively amid conditions of high uncertainty. ### Actor An actor is anyone or any entity that plays a role and has agency in a group, community, collective, network or ecosystem. This can refer to an individual person, organizations (like businesses, community groups, or government institutions), or other social groups. Think of actors as the "players" who can act, make decisions, affect others, and create momentum for change, triggering effects across their community or ecosystem and beyond.

Figure 65: The adjacent possible: from what is known to what becomes possible through new connections.[1]

### Adjacent possible The adjacent possible is a "shadow future"[\[2\]](#footnote-1) of possibilities that are already there, but have not been discovered yet[\[3\]](#footnote-2) (Figure 65). It is a serendipitous space of unseen ideas and solutions. Remixing these ideas with what we already have, can form a bricolage that catalyses R\&D. To unlock the possibilities of this space, we must intentionally explore what comes next and look beyond our current knowledge. We need to do this continuously because the adjacent possible is an infinite space that keeps expanding: each time we find something new or make a new connection, we unlock more possibilities that weren't available before.[\[4\]](#footnote-3) While the adjacent possible is bounded by what's achievable from our current state, our curiosity determines how far and wide we explore within these boundaries.[\[5\]](#footnote-4) ### Agency Agency[\[6\]](#footnote-5) is the capacity of actors to act, having the autonomy to make decisions and influence their circumstances to produce a certain effect or change in the world. Innovation, as such, is essentially about exercising agency, with the intent to bring change to the world. It operates at multiple levels: individual agency enables people to influence their immediate circumstances and conditions, while collective agency emerges when actors combine their energy and expertise to create broader change in their ecosystems. In any given ecosystem, agency is often unevenly distributed; some individuals or groups have better relationships, more power, greater resources, and more freedom to act than others. Agency can be temporary or situational; it can shift over time and across contexts. An actor might have agency in one situation but face barriers like structural inequalities, discrimination, or lack of access that limit their agency in another. By understanding who possesses agency and what blocks it, we can identify whom to involve in initiatives and whom to empower. ### Agile Agile is an iterative approach to software development, innovation and R\&D that emphasizes flexibility, continuous improvement, and rapid adaptation to change. We work in short cycles, adjusting course based on our insights gained from experiments, explorations, system interventions, and engagement with the ecosystem. ### Ambiguity Ambiguity refers to a situation in which multiple interpretations of the same information are possible, making it difficult to establish a single, clear meaning or direction.[\[7\]](#footnote-6) We experience ambiguity, for example, when we face multiple possible meanings that could all be valid simultaneously, when we encounter more questions than answers, or when the following steps are unclear. In complex systems, ambiguity emerges from the diverse perspectives, values, and mental models that stakeholders bring to the same challenge. Embracing ambiguity and its plural qualities, rather than reducing, eliminating, or ignoring it, opens space for creative tensions that may feel confusing at first but can reveal new possibilities.[\[8\]](#footnote-7) ### Assumption An assumption is a belief or fact that is considered to be true without direct proof to warrant claims. In development work, assumptions often underlie our inferences about what causes problems and how interventions will generate outcomes.[\[9\]](#footnote-8) Individuals and groups may not be aware of their assumptions – these hidden beliefs can trick the mind. Therefore, it's important to be critical and interrogate what we know and why something is considered to be true.[\[10\]](#footnote-9) Innovation processes often involve activities, methods and tools that help different actors in a system make their assumptions explicit. Testing ideas, directly engaging with key stakeholders, and observing them in their everyday context helps challenge our assumptions. ### Bias Biases are thinking errors that individuals, teams, and groups make when processing information, affecting the decisions and judgments they make. Some common examples include Confirmation Bias (cherry-picking information that confirms existing beliefs), Functional Fixedness (being blind to alternatives), Group Think (prioritizing consensus while avoiding controversy), Spotlight Effect (searching only where information is easily found), and Not Invented Here (favoring internal over external solutions). We're often unaware of these biases operating in our thinking.[\[11\]](#footnote-10) Many innovation methods, both explicitly and implicitly, aim to help identify, unpack, and challenge these biases.[\[12\]](#footnote-11) ### Bottom up "Bottom up" means we start at the local, community level rather than beginning with top-down policies or global level. This approach helps us unearth and promote local knowledge and experiences from people who are most affected by an issue. Working this way reveals solutions – and unaddressed needs – that are deeply rooted in local context and lived experience, showing us how communities and grassroots innovators are adapting to new uncertainties around drought, waste management, food security, digitalisation, disinformation, energy, mobility, etc. By scanning for these bottom-up signals across communities, we can identify emerging global patterns in sustainable development.[\[13\]](#footnote-12) ### Bricolage Bricolage is a creative act of combining or reconfiguring available resources, technologies and solutions for novel or unexpected purposes. Bricolage is about making the most of what’s already there. Since Claude Lévi-Strauss described bricolage as an innate human capability,[\[14\]](#footnote-13) it has been used and studied in various fields (entrepreneurship, psychology, organisational development).[\[15\]](#footnote-14) In our approach to R\&D, we use bricolage to develop frugal solutions for sustainable development, solving problems with readily available materials in resource-limited contexts. While bricolage often happens naturally as we solve problems,[\[16\]](#footnote-15) consciously looking for other potential uses or functions of a given object or solution, by framing the perceived affordances[\[17\]](#footnote-16) differently, expands our space of possibilities.[\[18\]](#footnote-17) Particularly in the adjacent possible, where we look for yet undiscovered solutions, bricolage helps us recognise the potential of these solutions, connect them with known solutions, and test new configurations[\[19\]](#footnote-18) that help us learn how they address emerging needs or generate value in unexpected ways. ### Catalyzing Catalyzing is an activity within our R\&D modes that focuses on increasing the rate or momentum with which an innovation scales through an ecosystem. It involves identifying and activating the right connections that can trigger broader change and amplify impact. Catalyzing happens when there's the alignment of needs or interests among different actors. They see that a solution or innovation can benefit everyone. When their interests align, adoption speeds up because stakeholders actively support and use the innovation to solve their own problems. These changes can happen unintentionally, e.g. through experiments, or intentionally, by carefully crafting specific relationships that allow information or knowledge to flow that eventually creates network effects of learning.
Figure 66: Pace layers of change: illustrating how different aspects of systems move at different speeds.[20]
### Change Change is the process or action that results in something becoming different from its previous state. Change is inherent to our work: we either intentionally try to create it, or it results in a situation we have to respond to. In our work, we often focus on change within complex systems, seeking new pathways for transformation. This can involve multiple aspects, ranging from reconfiguring core structures, relationships, and patterns of behavior that shape how a system functions, to shifting perspectives and developing new mental models to better align our understanding of the world with how it actually is. These different aspects of a system move at what Stewart Brand describes as different *pace layers of change*[\[21\]](#footnote-20) (Figure 66). For some aspects change can be sudden, while for others it can be slow, evolving over many years and perhaps even generations.[\[22\]](#footnote-21) ### Co-creation Co-creation[\[23\]](#footnote-22) is a collaborative approach where different stakeholders work together as equal partners to develop solutions, combining their unique perspectives, knowledge, and resources. This approach recognizes that the best answers often emerge from combining different types of expertise: the lived experience of communities, the technical knowledge of specialists, the perspectives and expertise of policy makers, and the know-how of people working directly on the ground. Unlike traditional approaches where experts hold decision-making power, co-creation distributes this power equally among participants. This fosters ownership of solutions, ensures better alignment between needs and capabilities, and helps reconcile different interests. ### Collective A collective is a group of actors who come together around a shared purpose and take coordinated action to achieve common goals. They are driven by a clear intention to create change or tackle specific challenges. Collectives can be temporary (like a task force addressing an urgent issue) or long-term (like an alliance working on systemic change). What makes them unique is their emphasis on joint action; members don't just share information or identity, they actively contribute to shared goals. They often establish clear agreements about how they'll work together, make decisions, and share responsibilities. The strength of a collective comes from its ability to mobilize diverse capabilities toward concrete outcomes and adapt as needed. While collectives share similarities with communities, networks, and ecosystems, yet each operates on different logic: networks connect, communities care,[\[24\]](#footnote-23) collectives act, and ecosystems catalyze. ### Collective intelligence Collective intelligence emerges when people work together, often supported by technology, to mobilize a wider range of information, ideas, and insights to address a specific issue.[\[25\]](#footnote-24) While humans have collaborated to solve problems since the beginning of time, what's new is our growing understanding of how collective intelligence works and how we can intentionally design for it. Also, today's digital technologies and ability to tap into diverse and real time data sources enables us to amplify collective intelligence at unprecedented scale and speed. For UNDP Accelerator Labs, collective intelligence design as a method, has been foundational. It shapes how we work by creating the conditions and tools that help groups become smarter together. This involves deliberately bringing together diverse perspectives, knowledge, and experiences while fostering an environment where collective learning can thrive. By designing for collective intelligence, we can better tackle complex sustainable development challenges[\[26\]](#footnote-25) that no single person or organization can solve in isolation. ### Combination Combination is the process of bringing together existing elements, ideas, technologies, or solutions to create something new that has different properties or capabilities than its individual parts. As Steven Johnson notes, "the secret to innovation is combining odds and ends"[\[27\]](#footnote-26) This approach is a fundamental mechanism for creating new solutions. The more odds and ends that are available for combination the greater the potential for unexpected connections and new innovations.[\[28\]](#footnote-27) In our R\&D practice, we intentionally expand the pool of odds and ends by mapping ecosystems to find the "oddballs"[\[29\]](#footnote-28) and grassroots innovators and start-ups to understand how they solve problems. We explore the adjacent possible to discover solutions and experiments that exist beyond our current awareness. When we design solutions, we create them to easily connect with other technologies and innovations, forming new bricolages that address unforeseen needs and unlock unexpected possibilities or functions. We achieve this through open-ended design, open licensing, standardized interfaces like APIs, and common data formats. This interoperability enables ecosystems to find new pathways to sustainable development. ### Commons Commons are shared resources governed by community-developed rules and norms where users collectively manage, maintain, and benefit from the resource.[\[30\]](#footnote-29) For our work, we have a specific interest in building an innovation commons for sustainable development.[\[31\]](#footnote-30) Such innovation commons are repositories of freely accessible, "open source" innovation-related tools, insights, solutions, data, and practices that significantly benefit innovators and other ecosystem actors.[\[32\]](#footnote-31) By pooling these resources and making them accessible and usable, more people, communities, and institutions can engage in innovation. This reduces the cost of innovation since participants can build upon what's already available rather than starting from scratch. Furthermore, innovation commons enable the ad hoc formation of groups or collectives around specific challenges or opportunities, bringing together diverse perspectives and expertise.[\[33\]](#footnote-32) The open nature of the commons facilitates the discovery of new opportunities[\[34\]](#footnote-33) or allows others to pursue the leads we uncover. This process creates new bricolages by combining ideas, insights, and innovations in new ways, reconfiguring and enhancing them, or taking them to scale.[\[35\]](#footnote-34) ### Community A community is a group of actors who share a common identity, interest, practice, or location and build relationships based on trust, mutual support and care.[\[36\]](#footnote-35) They form around a shared sense of belonging, where members support each other beyond immediate practical needs. Communities can be place-based (like neighborhood groups), interest-based (like professional communities), or practice-based (like groups sharing specific skills). Communities develop through regular interaction between members, creating environments where people share knowledge and experiences openly, learn from each other, and build lasting connections. While communities share similarities with networks, collectives, and ecosystems, each operates on different logic: networks connect, communities care,[\[37\]](#footnote-36) collectives act, and ecosystems catalyze. ### Condition Conditions are the circumstances or factors that enable or constrain how a system functions and evolves. They determine what's possible within a system. Some conditions can be influenced directly, like physical conditions (available space, infrastructure), social conditions (relationships, trust levels), or institutional conditions (policies, procedures). Others lie beyond immediate control, such as cultural conditions (beliefs, values), economic conditions (market dynamics, resource distribution), power dynamics (decision-making authority, influence), or environmental conditions (climate, geography). Creating change requires understanding both the conditions that can be shaped or amplified, and those that must be worked within. Through mapping and experimenting with these conditions, it becomes clear which ones to amplify or dampen to influence how the system evolves. Conditions aren't fixed; they interact with each other and change over time, creating new possibilities or constraints for system transformation. ### Configuration Configuration refers to how different elements in a system are arranged[\[38\]](#footnote-37) and connected. For example, from the physical layout of urban spaces to the formal and informal structures within organizations. These arrangements shape how systems function and the effects they produce. We map the system to understand how elements are configured and where to intervene. Through experimentation, we reconfigure how elements (i.e. actors, data sources, technology solutions) are connected and interacting, creating new patterns in a system’s defining relationships.[\[39\]](#footnote-38) We observe what emerges from these new configurations to see whether their effects are desirable or not, amplifying the patterns that work and dampening the ones that don't.[\[40\]](#footnote-39) ### Confusion Confusion is inherent to uncertainty. It is the result of "not knowing", it emerges when our existing mental models, knowledge and expectations no longer match with reality. We can distinguish between bad confusion, which leads to inertia or poor decisions, and good confusion, which drives learning and builds resilience. In collective action and learning the challenge is to avoid bad confusion while nurturing good confusion by focussing our attention on what we're curious about; it allows us to embrace ambiguity and uncertainty, rather than being overwhelmed by it. ### Culture Culture encompasses the shared beliefs, values,[\[41\]](#footnote-40) practices, symbols,[\[42\]](#footnote-41) worldviews, and ways of life that shape how communities understand and interact with their world. It includes observable aspects like rituals, traditions, and behavioral practices, as well as hidden dimensions such as collective memory, spiritual beliefs, knowledge systems, cosmovisions,[\[43\]](#footnote-42) and perceptions of time and space.[\[44\]](#footnote-43) Culture is dynamic, continuously evolving through internal innovations and external influences, while maintaining core elements that provide identity and continuity across generations. ### Curation Curation is the intentional process of nurturing networks and collectives by forging connections between actors and establishing conditions that facilitate the flow of knowledge and catalyze collective learning and action.[\[45\]](#footnote-44) Curation makes a network or collective worth connecting with, and staying engaged. While it's often assumed that great talent and knowledge will naturally curate themselves, intentional curation is essential for networks and collectives to become more impactful than just the sum of their members. Unlike the art world's “hero-curator” model where a single curator determines the success of exhibitions or galleries, curation of networks and collectives is typically a shared responsibility undertaken by core teams or distributed across multiple members. In collective R\&D, effective curation is a multi-faceted process that involves specific responsibilities and capabilities. This includes crafting narratives that foster a sense of shared purpose and belonging, as well as stewarding membership and onboarding newcomers by helping them appreciate the – often implicit – values and norms of the network. It involves actively connecting people through brokering relationships and organizing events or activities that strengthen those connections while also creating opportunities for serendipitous encounters. To unlock its collective wisdom, curation involves coordinating the curiosity and collective learning objectives of a network by framing shared learning questions and making knowledge accessible and usable. ### Curiosity Curiosity is our drive to seek out and connect information, creating and expanding networks of knowledge[\[46\]](#footnote-45) and possibility.[\[47\]](#footnote-46) It fuels our learning by driving us to ask questions that uncover new perspectives and insights, and help us see the gaps in our knowledge. In R\&D, curiosity drives us into uncharted territory, enabling us to deepen or expand our knowledge while increasing the likelihood of making unexpected discoveries in the adjacent possible.[\[48\]](#footnote-47) This aligns with three information-seeking strategies described by Dani Bassett and Perry Zurn:[\[49\]](#footnote-48) A goal-oriented approach focuses on deepening knowledge around specific problems or questions. An associative approach follows connections between intersecting and adjacent domains, expanding our knowledge base through creative linkages. A serendipitous approach explores based on what seems interesting, leading to unexpected discoveries across diverse topics. Being aware of which approach we often use or naturally gravitate towards – and when to shift between them – helps us be more intentional in our learning and discovering, making our learning more effective. ### Data Data is the smallest unit of information that is collected and organized in a structured way to be analyzed, interpreted, or used for making sense of the world around us and decision-making. Data generally represents a variable, feature, or attribute (e.g. name, age) of a real-life entity (e.g. person, boat).[\[50\]](#footnote-49) Data comes from a wide variety of sources. We often distinguish between three main categories: traditional or official sources (such as census data, economic indicators or national statistics); non-traditional sources (like satellite imagery, social media, citizen-generated data, or sensor networks);[\[51\]](#footnote-50) and qualitative research data (such as ethnographic observations, interviews, and field notes). We work with various types of data, such as crowdsourced data (collected from public contributions), real-time data (continuously updated), open data (freely available for public use), and big data (extremely large datasets that require specialized tools to process).[\[52\]](#footnote-51) Our approach to data emphasizes empowerment rather than extraction, ensuring communities can access, understand, and benefit from the insights their data reveals, placing knowledge back in the hands of those who generate it.[\[53\]](#footnote-52) Data science is critical for our R\&D practice and enabling collective intelligence. When using large data sets or combining different data sources, data science helps us discover patterns, extract actionable insights, find positive deviants, or create a real-time understanding of a situation. ### Developing Developing is an activity within our R\&D modes that focuses on deepening our understanding of systems and identifying opportunities for change. It involves defining and reframing problems based on how systems work, while refining ideas, solutions, prototypes and interventions together with key players in the ecosystem. ### Development intelligence Development intelligence is the collection, analysis, and application of information that provides actionable insights for decision-making and advancing sustainable development. It integrates diverse sources of knowledge[\[54\]](#footnote-53) and data to enhance our understanding of complex development contexts.[\[55\]](#footnote-54) This intelligence draws from various sources, including evidence from experiments, ethnographic insights, grassroots innovations, local and traditional knowledge, signals of change, scientific research, policy analysis, non-traditional as well as official data, crowdsourced information, and system mappings that reveal the drivers behind development challenges and opportunities for positive change. By combining these different inputs, development intelligence gives us a more comprehensive real-time understanding of what's happening in local, regional and global development ecosystems. It helps us see development issues from multiple angles and levels, spot where new challenges are emerging, identify our knowledge gaps, and understand where we need to focus our learning efforts to accelerate progress towards sustainable development. ### Diffusing Diffusing is an activity within our R\&D modes that focuses on spreading knowledge, insights and solutions through a community or ecosystem.[\[56\]](#footnote-55) It involves making innovations accessible and usable, enabling others to build upon these ideas. Through diffusing, we create pathways for solutions to spread across different contexts and communities, so others can adopt, adapt and enhance[\[57\]](#footnote-56) them to meet their specific needs. ### Digital Digital refers to the use of computer technology, data, and connected systems to transform how we work, communicate, and solve problems. It encompasses both the technical aspects (like converting analog to digital formats) and the cultural shifts that come from working with digital technologies, from data analytics and automation to artificial intelligence and digital collaboration tools. ### Directed improvisation Directed improvisation is a concept we used when setting up the Accelerator Lab network, to provide our initiative with a clear purpose while making space for unexpected discoveries and adapting to unanticipated situations. The term comes from Yuen Yuen Ang's[\[58\]](#footnote-57) research on China's development and explains how the country achieved rapid growth. Governing a nation differs in scale and complexity from curating a globally distributed network of labs, but we found this idea useful to give the Labs guidance.[\[59\]](#footnote-58) While *directed* and *improvisation* might seem contradictory, in concert they enabled us to engage with the messy adjacent possible and create new pathways to sustainable impact in uncertainty. ### Directionality Directionality is setting a course toward desired change without fixing rigid end goals. In uncertainty, we can't predict exact outcomes, so instead of chasing predefined targets, we identify the general direction – a North star we want to move towards. Directionality is about starting from where we are,[\[60\]](#footnote-59) not where we think we should be. Think of it as "crossing the river while feeling for the stones," as the Chinese proverb goes. The destination may be clear, but each step depends on what's discovered along the way. This is reflected in our R\&D approach. While we keep a general sense of direction, we're open to making serendipitous discoveries in the field, gaining insights from experiments, developing new learning questions, etc., that help us move forward. ### Ecosystem An ecosystem is a dynamic network of diverse actors who are interconnected and interdependent, working together to create value and impact. Ecosystems can include organizations, groups, and individuals from various sectors, like businesses, government agencies, research institutions, and community organizations, each playing distinct but complementary roles. Ecosystems can also focus on a specific activity, like natural, biological ecosystems,[\[61\]](#footnote-60) or organisational ecosystems like business,[\[62\]](#footnote-61) or innovation[\[63\]](#footnote-62) ecosystems. These focus areas serve as a lens that helps us see and understand how the ecosystem is organized, why actors interact in a certain way, and how value is created. What makes ecosystems unique is that no single actor controls the whole system;[\[64\]](#footnote-63) instead, value emerges from how different actors interact, adapt, and combine or complement their capabilities,[\[65\]](#footnote-64) forming a symbiotic relationship.[\[66\]](#footnote-65) The strength of an ecosystem comes from how it enables actors to discover unexpected opportunities and create new value by combining their diverse capabilities; these processes of discovery and synthesis often happen in the "middle ground",[\[67\]](#footnote-66) the R\&D space that sits between institutional and grassroots initiatives. By orchestrating the exchange of ideas and innovations in this space and enabling collaboration, ecosystems enable innovations to scale and catalyze R\&D.[\[68\]](#footnote-67) Adopting an ecosystem approach is critical for addressing complex societal and development challenges that no single organization or sector can solve alone.[\[69\]](#footnote-68) While ecosystems share similarities with networks, communities, and collectives, each operates on different logic: networks connect, communities care,[\[70\]](#footnote-69) collectives act, and ecosystems catalyze. ### Effect Effects are the emergent outcomes that complex systems produce through the interactions between their parts and their overal configuration.[\[71\]](#footnote-70) These outcomes can show up as new behaviors like changed waste practices, adoption of new technologies, new relationships between local actors, or broader changes such as emerging markets. Effects can be positive or negative, depending on who experiences them and in what context. In complex systems, cause and effect relationships are non-linear and networked; there isn't always a direct link between interventions and their outcomes. Effects may emerge quickly or take time to appear, and can be felt either directly or far from their origin point. While some effects can be anticipated, others are unexpected. This means we need to look beyond just the intended and expected outcomes to understand and work with both (un)desired and unanticipated results as they emerge.[\[72\]](#footnote-71) ### Emergence Emergence is a feature of complex systems where new patterns, behaviors, or properties arise unexpectedly from the interactions between parts of the system. These emergent features cannot be predicted by understanding the individual parts alone, but rather result from how these parts relate to and influence each other over time.[\[73\]](#footnote-72) Traditional approaches to sustainable development often depend on careful planning, mapping out activities based on assumptions about how the world operates, with the belief that outcomes can be controlled and predicted.[\[74\]](#footnote-73) However, when we recognize the complexity of sustainable development challenges, we adopt different assumptions: we understand that the world is unpredictable and that outcomes cannot be controlled or anticipated. Instead of expecting predetermined results, we intervene and observe how the system responds. What effects emerge? Are they desirable – and for whom? Sometimes the outcomes align with our expectations, but often the system surprises us. ### Empathy Empathy is the ability and the willingness to understand and share the feelings, perspectives, and experiences of others by putting yourself in their position.[\[75\]](#footnote-74) It goes beyond just observing or analyzing their situation to grasp their emotions, challenges, and lived experiences. In R\&D, empathy helps us move past assumptions by experiencing people's daily realities to better understand what matters to them. ### Evidence Evidence is the data, information and facts that support or challenge our understanding of how the world – or a specific system – works. In R\&D, we create evidence through experiments to understand what works, where, why and for whom, as well as what doesn't work.[\[76\]](#footnote-75) By collecting and analyzing different types of data, we can make better-informed decisions and take action by understanding both the magnitude and urgency of an issue. Evidence provides legitimacy to our insights, solutions and innovations, enabling their adoption and implementation. It helps our initiatives move forward. ### Experimentation An experiment is a structured process that helps us learn what works and what doesn't. Experiments create evidence that informs our decisions and helps build legitimacy for an idea or innovation. There are many ways to experiment, from quick and simple tests to more rigorous trials that require more time and resources.[\[77\]](#footnote-76) A successful experiment is one we can learn from;[\[78\]](#footnote-77) it is an intelligent way to fail[\[79\]](#footnote-78) and it does that ideally for a low cost.[\[80\]](#footnote-79) In R\&D, experimentation creates evidence that informs our next steps and gives ideas legitimacy, generating stakeholder interest to take them forward. Experiments can also catalyze R\&D in unexpected ways by making tacit needs explicit; when stakeholders see a prototype, they often recognize needs they couldn't articulate before; they can only tell what they need when they see it. We never experiment alone, but actively involve and co-design with diverse stakeholders, from grassroots innovators to government agencies. This collaborative approach helps create ownership of both the intervention and its results; results carry further when people own them. ### Exploring Exploring is an activity within our R\&D modes that focuses on mapping what's already there, and finding the new in an ecosystem. By mapping existing solutions and knowledge, we're able to accelerate our learning and build on what works. Through exploring the edge of knowledge, we discover new possibilities, inputs and perspectives that help us move forward. ### Facilitation Facilitation is the practice of designing and leading group processes that unlock collective intelligence, create purposeful collaboration, and enable co-creation. As Anna Birney put it, facilitation is “a creative process of bringing awareness to the world.”[\[81\]](#footnote-80) As such, the role of a facilitator is to hold the intent of the room and channel the dynamic energy of a group in pursuit of that intent.[\[82\]](#footnote-81) Facilitation can take place in various settings, whether in person, such as in a meeting room or under a mango tree in the field,[\[83\]](#footnote-82) or online. In R\&D, effective facilitation promotes psychological safety,[\[84\]](#footnote-83) creating an environment where participants feel comfortable contributing openly and authentically. Psychological safety allows groups to engage with diverse, potentially conflicting perspectives[\[85\]](#footnote-84) and creates space for intentional ambiguity. Skilled facilitators balance structure with flexibility. They keep the process on track, manage the pacing of activities to align with energy levels, and remain sensitive to the group's dynamics, allowing emerging directions to unfold. When done well, facilitation often goes unnoticed, enabling participants to focus on their collaborative efforts rather than on the facilitation process itself. Good facilitation can transform a group of individuals into a collective that can generate insights and solutions beyond what any single person could produce alone. ### Failure Failure is an outcome of our decisions and actions that doesn't meet intended goals or expectations. While often seen as negative, Amy Edmondson[\[86\]](#footnote-85) helps us distinguish between two types: blameworthy and praiseworthy failure. Blameworthy failures are preventable mistakes that happen in predictable conditions; like violating protocols or being inattentive. Praiseworthy failures come in two forms: unavoidable failures that occur in complex or uncertain situations, and intelligent failures that happen when we intentionally push the boundaries of what we know through experiments. To find where the edge of knowledge lies, a percentage of experiments must actually fail.[\[87\]](#footnote-86) Failure, as such, is an inherent part of R\&D; it reveals boundaries, challenges assumptions, and produces critical knowledge that helps us find new pathways to impact. The value of failure lies in what we learn from it, whether intentional or not. This learning requires psychological safety, making it essential to build a culture where teams feel secure to discuss, learn from, and even celebrate failure.[\[88\]](#footnote-87) ![Figure 67: Framework in action: collectively developing an R\&D agenda for digital financial inclusion (Nairobi, June 2024)](/files/6b3bfe3424c3f3f78b676c45964f4e6772de8ddb) ### Framework A framework is a representation of an individual or collective mental model that helps us make sense of complex situations and guides our actions. Often, frameworks are presented in a visual form,[\[89\]](#footnote-88) ranging from rough sketches to sophisticated diagrams, making tacit mental models explicit and turning individual insights into collective understanding. Visualization allows us to emphasize key elements and relationships, organizing concepts, principles, activities, methods, and other entities while explaining the underlying logic or narrative. In R\&D, frameworks can take many forms and serve various purposes, including documenting research findings,[\[90\]](#footnote-89) serving as blueprints for solution architectures, providing structure to processes[\[91\]](#footnote-90) (Figure 67), and offering guidance for strategy and decision-making.[\[92\]](#footnote-91) Most importantly, frameworks facilitate interactions among individuals coming from different social worlds.[\[93\]](#footnote-92) Frameworks that provide new perspectives or tools for new thinking can travel far, inspiring many people in unexpected places. These frameworks can spread like memes[\[94\]](#footnote-93) from a core team to other actors within an ecosystem, accelerating diffusion across an ecosystem. ### Frugality Frugality means doing more with less. It often comes from the need to develop solutions in situations where resources are limited.[\[95\]](#footnote-94) In sustainable development, frugal approaches leverage readily available materials and indigenous knowledge to address challenges in ways that fit the local context. Frugality and simple, low-tech solutions are key to grassroots innovation. Local innovators use creative and resourceful methods to solve problems in their communities.[\[96\]](#footnote-95) These solutions tend to be multifunctional, addressing several needs simultaneously.[\[97\]](#footnote-96) Frugal solutions also offer greater flexibility; their simpler designs, lower costs, and independence from complex resources allow them to be more easily adapted, modified, or repurposed as needs change.[\[98\]](#footnote-97) ### Future The future refers to the state of the world that will or is likely to happen in time, both in the near term and long term. There isn't just one future, but rather a spectrum of multiple possible futures: some more plausible, others more probable, and certain ones more preferred.[\[99\]](#footnote-98) To explore and understand what these futures might look like, we scan horizons to identify weak signals of change, we reflect on their possible implications and develop scenarios about how the future might unfold and what it might look like for different actors.[\[100\]](#footnote-99) Working with futures and using foresight methods isn't about predicting what will happen; it's about anticipating possibilities so we can better prepare for what may come, whether they are likely or unlikely. In our work, we don't consider a future as something that simply happens to us; it's something we can collectively imagine[\[101\]](#footnote-100) and actively shape together.

Figure 68: Three approaches to innovation: designing for people (design-led), designing with people (co-creation), and innovation by people (grassroots-led). Each approach reflects different roles and relationships between those who design solutions and those affected by an issue.[102]

### Grassroot innovation Grassroots innovations are often local, home-grown solutions led by women and men (Figure 68) who innovate to address their needs with limited resources.[\[103\]](#footnote-102) These solutions may never have been codified or disseminated beyond their communities, nor scaled; these are often unglamorous innovations done by non-elites.[\[104\]](#footnote-103) At the UNDP Accelerator Labs, we have been exploring and documenting grassroots innovation since day one.[\[105\]](#footnote-104) Grassroots innovators have a unique kind of knowledge as they are closest to the problem[\[106\]](#footnote-105) and often draw on traditional knowledge, frugal problem-solving and practical creativity[\[107\]](#footnote-106) – they work with what they have.[\[108\]](#footnote-107) Once we start to see the ingenuity of these grassroots innovators, we cannot unsee it. Examples are countless and address a wide variety of needs: farmers developing dual-purpose water turbines generating electricity while irrigating fields,[\[109\]](#footnote-108) communities using bamboo and coconut shells as nature-based solutions for coastal protection and flood prevention,[\[110\]](#footnote-109) or using insects to address waste issues, food insecurity, or plastic pollution,[\[111\]](#footnote-110) to name just a very few examples. Most grassroots innovations seem to be multifunctional, designed to solve multiple problems, not just one.[\[112\]](#footnote-111) In our R\&D approach, grassroots innovations provide valuable insights from real-life experiences. They help us identify unaddressed needs and learn from field-tested solutions.[\[113\]](#footnote-112) They represent the adjacent possible: solutions that already exist but remain undiscovered or underappreciated by conventional institutions. These innovations often demonstrate approaches that better fit local needs, offering more sustainable, resource-efficient, and culturally appropriate pathways. ### Hypothesis A hypothesis is a testable statement that describes what we expect will happen when we take specific actions. A hypothesis is not just an idea: instead a hypothesis helps us test good ideas – because ideas are only as good as they sound, until they are tested.[\[114\]](#footnote-113) When developing these statements, we consider what needs to happen for an intervention to work, what effects we expect to see, and how we'll measure whether we're right or wrong. The most common structure is an if-then statement: "if we do this, then we expect that to happen". This approach gives rigor to our experiments and helps structure our learning by making explicit what we believe might work and why. Unlike scientific research that seeks fundamental truths, in R\&D we use hypotheses to test ideas and interventions that can create real value and catalyze broader system change. ### Idea An idea is a belief about how a specific arrangement of actors, solutions, information and technology might create certain effects. They are “works of bricolage” as Steven Johnson[\[115\]](#footnote-114) put it. But ideas are only as good as they sound until they face reality[\[116\]](#footnote-115) – rarely do they survive their first encounters intact. That's why we turn ideas into testable hypotheses, experimenting in the real world, developing and improving[\[117\]](#footnote-116) them to discover what effects they actually create and whether these are desirable. ### Inclusivity Inclusivity in innovation practice refers to new products, processes, or approaches that strive to improve the lives and livelihoods of those directly affected by challenges, as well as marginalized and excluded communities.[\[118\]](#footnote-117) Inclusive innovation prioritizes social objectives and considers the local context throughout the process, emphasizing that solutions should be created with communities rather than imposed on them. We need to acknowledge that innovation is not neutral; it has both pace and direction.[\[119\]](#footnote-118) Every innovation approach has an underlying logic (often implicit) that can either amplify equality and social outcomes or exacerbate inequalities. For R\&D in sustainable development, inclusivity ensures that our solutions address real needs, build on local knowledge, and create outcomes that are equitable, effective, and lasting. ### Innovation Innovation involves coming up with new ideas that are successfully implemented to create value for people, communities or society.[\[120\]](#footnote-119) Often these ideas aren't entirely new – instead, they combine or reconfigure existing solutions[\[121\]](#footnote-120) in fresh ways to address needs or deliver better value. For an innovation to succeed, we need to test, develop and improve these ideas until they create real benefits for people.[\[122\]](#footnote-121) The key is turning promising concepts into practical solutions that work in the real world. ### Insight An insight is a deep understanding or realization that reveals patterns, relationships, or opportunities in a situation or challenge that were not seen before. Insights emerge from carefully analyzing data, observations, and experiences, often connecting different pieces of information in new ways. They help us better understand what is happening, inform decisions and create new possibilities for action. For our global network, insights represent generalizable learnings and patterns we discover across multiple countries and contexts. These insights may inform our work elsewhere and help us codify our learnings into value propositions.[\[123\]](#footnote-122) ### Issue An issue[\[124\]](#footnote-123) is a persistent problem within a system that produces adverse effects and where current approaches are insufficient. Issues emerge from complex situations, conditions, or interconnected factors that present both difficulties and opportunities for intervention. In terms of difficulty, issues generally fall into two categories: technical problems and adaptive challenges – often also referred to as tame and wicked problems.[\[125\]](#footnote-124) Technical issues can be clearly defined and broken down into manageable parts (like assembling a car).[\[126\]](#footnote-125) They have known solutions that can be addressed by applying existing know-how and established methods (i.e., best practices). Technical issues assume a stable, ordered world with predictable outcomes and a clear end-state; once it’s solved, it stays solved. They usually require technical or authoritative expertise to solve. Adaptive challenges, by contrast, resist clear definition and are often caused by people's beliefs, habits, and behaviors (such as school-run traffic congestion) and conflicting interests. They have no precedents or silver bullets and require exploration, experimentation, and sense-making to improve the situation. Adaptive challenges recognize a dynamic, sometimes chaotic world with unpredictable outcomes and possibly no end-state.[\[127\]](#footnote-126) Addressing them demands an ecosystem approach that leverages diverse perspectives and the collective intelligence of stakeholders. Treating adaptive challenges as technical ones may seem appealing and yield some short-term improvements, but can exacerbate the issue over time. Also, development challenges are not always clearly labeled as "technical" or "adaptive"; rather, they present themselves as a mix of interconnected technical and adaptive issues.[\[128\]](#footnote-127) Our R\&D approach focuses on identifying emerging development challenges that have a critical impact on large and vulnerable populations. They are not owned by any single stakeholder and require collective intelligence and action to address. We explore the ecosystem and map stakeholders. We form collectives by initiating partnerships where there is momentum and shared interest to work together. Through collaborative experimentation and learning, we navigate complexity to find pathways for change and impact. ### Lab A lab refers to a dedicated team, unit, or function that employs exploratory and experimental methods to address social, public or business challenges and to create new forms of impact and value.[\[129\]](#footnote-128) They serve as safe spaces with a clear mandate to "figure things out", that allows them to explore, test, and develop new ways of working, collaborating with unusual ecosystem actors, using new methods and technology. These labs vary in size and form, they most commonly take the form of temporary projects or more permanent structures. They can sit inside or outside of an organisation, or right in between, being closely connected to everyday business or have a more future-oriented outlook. They may focus on specific domains like healthcare, education, or circularity, or work broadly across sectors where they identify unaddressed needs, unused potential or emerging risks. Depending on these challenges, their approach can range from focusing on one or two specific innovation methods to employing a broader and dynamic palette. ### Learning Learning is the process of gaining new understanding and knowledge about the world around us.[\[130\]](#footnote-129) From the start, learning has been the modus operandi of the Accelerator Labs; they were specifically mandated to accelerate learning to keep up with the pace of change and close the gap between current results and bigger aspirations for sustainable development. As Toby Lowe[\[131\]](#footnote-130) and colleagues remind us: in complexity, learning is the strategy to achieve better outcomes. We learn by observing what's happening in communities, convening meetings with people who bring new perspectives, actively exploring new situations, collecting and analyzing data, experimenting to find out what works, and reflecting on experiences[\[132\]](#footnote-131) and our thinking.[\[133\]](#footnote-132) Please note, often, learning isn't just about acquiring new knowledge or adopting new thinking but also unlearning the outdated mental models and habits of mind that no longer serve us.[\[134\]](#footnote-133) In R\&D, learning happens both individually and collectively; when we bring together different perspectives and experiences, we create deeper understanding of problems and potential solutions. This collective learning helps us challenge our assumptions, discover unexpected opportunities, and develop more effective approaches to sustainable development. ### Learning circle Learning circles involve facilitated dialogue among knowledge holders or experts[\[135\]](#footnote-134), who come together to discuss a specific learning question.[\[136\]](#footnote-135) With learning circles we explore aspects of a development challenge where we need to accelerate our knowledge and understanding. In these conversations, we involve actors with diverse backgrounds, perspectives, expertise and inputs to rapidly identify the edge of what is known and collectively become smarter on a topic. Learning circles aim to unlock tacit knowledge and instigate knowledge flows across participants. To elevate the tacit knowledge[\[137\]](#footnote-136), we let participants talk about their experiences, observations, what they learned from changing or intervening in systems, running experiments, etc. As Dave Snowden notes, "We know more than we can tell, and we tell more than we can write"[\[138\]](#footnote-137) – which is why we focus on talking. ### Learning cycle Learning cycles was the initial framework for our process when we launched the Accelerator Lab Network. It consisted of four stages: sense, explore, test and grow; which later evolved into our framework of R\&D modes. Each cycle was meant to take 100 days, with Labs developing Learning Plans focused on specific Learning Questions. Reality proved messier and didn't always fit this 100-day rhythm, so we let go of that timeline. However, the initial pressure to complete short cycles helped us get started by taking action and learning from experience rather than overthinking it and creating a perfect plan. Over time Labs have developed their own frameworks to structure their processes. Learning Plans still remain an important tool, serving as one of our key data sources to track what the Accelerator Lab Network is working on and learning. ### Legitimacy Legitimacy[\[139\]](#footnote-138) is having a social license to operate. For R\&D it means that key stakeholders accept an initiative, approach or idea as something relevant, appropriate and valued for their community, collective or organization. Legitimacy helps to get a mandate: to try out new ways of working, challenge the status quo by asking different questions or introducing new perspectives, collaborate with unconventional partners, get resources for exploration or experimentation even if there is a chance of failure or no clear outcomes. Given the collaborative and collective nature of R\&D, creating and maintaining legitimacy is critical and an ongoing process.[\[140\]](#footnote-139) ### Leverage point A leverage point is a place in a system where a small change can produce big effects across the whole system. When working on complex challenges, it's crucial to identify these strategic points where focused interventions can trigger broader positive changes. As Donella Meadows put it: “leverage points are points of power”.[\[141\]](#footnote-140) They can take many forms; from changing rules and incentives, to shifting information flows, to transforming deeply held beliefs. While high-level leverage points (like changing mindsets or goals) can create more fundamental change, they are often harder to influence than more practical leverage points (like adjusting feedback loops or changing how resources flow).[\[142\]](#footnote-141) Through careful observation and system mapping, we can discover these powerful intervention points and use them strategically to create lasting change. ### Local knowledge Local knowledge is the knowledge that people in a given community have developed over time, and continue to develop.[\[143\]](#footnote-142) Local knowledge encompasses many vital areas of community life; from food security and healthcare to agriculture, natural resource management, environmental conservation, climate resilience[\[144\]](#footnote-143) and social organization. It is rooted in lived experience and embedded in the cultural practices, relationships, and daily life of communities. Local knowledge enables communities to understand and respond to challenges by combining time-tested wisdom with new approaches that fit their specific context and values, while preserving their cultural identity. Local knowledge exists in both documented and undocumented forms; while some aspects may be written down or recorded, much of it is passed on across generations through practice, storytelling, and rituals, making it essential to engage directly with communities to understand and learn from their knowledge. Because of its long history, it is often referred to as ancestral, traditional, indigenous or grassroot knowledge.[\[145\]](#footnote-144) While terms like *ancestral* and *traditional* knowledge might imply something static, the nature of these knowledge systems is in fact dynamic and continuously evaluated, renewed and adapted to *l'esprit du temps*[\[146\]](#footnote-145) and its challenges. Local knowledge is therefore a rich source of wisdom about how local systems work[\[147\]](#footnote-146) and evolve. As a global learning network, we identify patterns across these local systems and elevate local knowledge to understand what communities worldwide are grappling with and to inform our global R\&D agenda.[\[148\]](#footnote-147) It should be noted that in our practice, we often observe what we call a “battle of epistemologies” where local, traditional, and indigenous knowledge systems appear incompatible with scientific and policy knowledge systems.[\[149\]](#footnote-148) Instead of treating these as separate or opposing systems, our R\&D approach seeks to create bridges between them. We leverage the unique strengths and insights of each knowledge system to enable collective intelligence and inform people's everyday decisions, communities planning their futures, governments developing policies, and more. ### Logic A logic describes a consistent way of reasoning and deciding that shapes how we see and act in the world. In R\&D, our logic determines how we learn, create knowledge, develop solutions, and create value. When collaborating with others, alignment between logics is essential for working together effectively. ### Mandate A mandate is the formal or informal permission to explore, experiment and try new approaches, even when outcomes are uncertain or unknown. Think of it as a license: the permission to step outside business-as-usual to test unconventional ideas, work with unusual partners, or address emerging issues that are outside the current scope of work. For R\&D teams and innovation labs, having a clear mandate is crucial both inside their organization and with external partners. Without internal mandates from leadership and external mandates from communities and stakeholders, R\&D efforts struggle to get the support and resources needed to succeed. ### Method A method is a structured approach with specific steps and activities that helps us achieve intended goals and outcomes. Methods give us guidance in our R\&D cycle; from learning about systems and their dynamics, to understanding problems, developing solutions, testing ideas, generating insights, and enabling collaboration, etc. Methods can range from simple techniques like interviewing and observation to more advanced approaches like systems mapping or participatory design.[\[150\]](#footnote-149) Each method draws on different skills and capabilities, serving specific purposes, but each has its limitations; that's why we often combine different methods in our work. Also note, we tend to favor methods from our own professional background or discipline, but it's important to look beyond these personal biases and consider diverse approaches that might better serve our goals.[\[151\]](#footnote-150) Figure 69: The middleground: An R&D space between formal institutions and grassroots innovators where knowledge, ideas, and innovations are exchanged and actors connect.[152] ### Middle ground In an ecosystem, the "middle ground"[\[153\]](#footnote-152) is an R\&D space that sits in between the formally structured institutions and the fluid and loosely organised grassroots (Figure 69). The institutional layer includes formal organizations like governments, multinational companies, large NGOs, development agencies, and academia. The grassroots layer involves local innovators, social entrepreneurs, start-ups, small NGOs and businesses. We often find ourselves operating in the space in between, the middleground, where we play the role of brokers[\[154\]](#footnote-153) and facilitators, enabling the exchange of ideas, knowledge and innovations between these institutional and grassroot actors. We do this by convening actors from across the ecosystem, hosting events and meetings, facilitating workshops, and developing platforms.[\[155\]](#footnote-154) These activities create opportunities for ideas, insights and innovations to connect and start to flow between the two constituencies. Figure 70: Mindsets are grounded in beliefs and worldviews that shape how we see, think, and act in the world. ### Mindset Mindsets involve a set of beliefs, attitudes, and worldviews that shape how we perceive, interpret, and engage with the world around us (Figure 70). In more simple terms, mindsets are "habits of mind"[\[156\]](#footnote-155) that shape how we see, think and act in the world.[\[157\]](#footnote-156) Seeing is about perceiving and paying attention to the world around us through the filters of our existing worldviews and mental models. Thinking involves making sense of our observations and experiences, both consciously and subconsciously,[\[158\]](#footnote-157) to form ideas and plan actions based on our beliefs and experiences. Acting refers to how we act upon our observations and thoughts, and how we engage with the world to affect change.[\[159\]](#footnote-158) Organizations often aspire to transform themselves by adopting new mindsets, like an innovation, digital, agile mindset[\[160\]](#footnote-159) to adapt to a changing world. However, actually changing mindsets is a long-term, multifaceted process – it is easier said than done. ### Mode Modes form the overarching framework of our R\&D practice, consisting of three distinct approaches: Sense & Explore, Develop & Test, and Diffuse & Catalyse. Each mode combines a specific way of doing with a way of being, shaping how we engage with the world, experience and learn from it. These modes help us direct our intentions, focus our attention, and guide our actions in contexts characterized by ambiguity, complexity and uncertainty. ### Momentum Momentum refers to the build-up of energy in a system that reinforces a direction of change. This energy manifests through patterns, behaviours and self-reinforcing feedback loops that emerge over time. We can see these feedback loops in action through growing political capital, increased agency, or greater willingness of actors to take action. But momentum requires constant attention; reinforcing patterns can weaken or change direction, causing a system to lose its drive for change. ### Network A network consists of actors[\[161\]](#footnote-160) who are connected and interact with each other to achieve common goals.[\[162\]](#footnote-161) For networks, the quality and quantity of relationships is critical to be effective.[\[163\]](#footnote-162) Through these connections, actors share knowledge, resources, and support, enabling the network to tackle challenges that individuals couldn't address alone. While some networks are formally organized with clear structures, others emerge and evolve through informal connections. Networks vary widely in their makeup; they can be small local groups or vast global systems connecting millions.The UNDP Accelerator Labs is an example of a global network, operating as what Liz Altman calls a learning “network of ecosystems”[\[164\]](#footnote-163) that connects local insights to global challenges. This network structure helps us learn faster by exchanging innovations, insights and practices across contexts, revealing patterns of emerging development challenges and value creation for sustainable development. While networks share similarities with communities, collectives, and ecosystems, each operates on its own logic: networks connect, communities care,[\[165\]](#footnote-164) collectives act, and ecosystems catalyze. ### Network effect Network effects[\[166\]](#footnote-165) occur when the value and knowledge of an ecosystem increases as more people are involved, actively contribute and use it. Each additional user contributes to and benefits from the overall value of the ecosystem, causing learning and innovation to scale exponentially; instead of linearly. In R\&D, we can leverage network effects by increasing the number of actors involved, increasing their diversity, finding or developing incentives to contribute, and creating meaningful connections across an ecosystem. The strength of these effects depends not just on the number of participants, but also on the quality and frequency of their interactions. ### Open Open refers to making our data, knowledge, experiments, tools and practice accessible and usable by anyone. It means being transparent about what we do and how we do it, while inviting others to build upon and improve our work. When others bring their perspectives and adapt our work to their contexts, they create new possibilities we couldn't have imagined. Through our global innovation commons for sustainable development,[\[167\]](#footnote-166) we aim to multiply impact beyond our immediate reach and remit.[\[168\]](#footnote-167) This helps solutions evolve to meet local needs while catalyzing innovation through unexpected combinations of what we openly share. ### Open-ended design Open-ended design is an approach where designers intentionally create "unfinished" solutions[\[169\]](#footnote-168) that users can adapt, hack and complete based on their specific needs and context.[\[170\]](#footnote-169) It recognizes people's agency and creativity as problem-solvers who can enhance and modify solutions to better fit their circumstances.[\[171\]](#footnote-170) By designing with intentional openness and flexibility, solutions can better connect with existing practices and other solutions in unexpected ways. This "loose fit"[\[172\]](#footnote-171) or “clumsiness”[\[173\]](#footnote-172) helps us – as designers – mitigate our biases, address assumptions and knowledge gaps about evolving needs, while creating space for creativity and serendipity. This approach can be applied not only to end-solutions, but also to prototypes earlier in the process, leveraging the ingenuity and deep knowledge people have of their own situation, but keep in mind open-ended design also depends on people having the capability and motivation to adapt and enhance existing solutions.

Figure 71: Managing paradoxes is like balancing a ball on a convex surface: using the tension between two opposing forces productively, even when holding both seems impossible.

### Paradox A paradox represents a tension between contradictory – yet related – ideas or approaches that must be maintained rather than resolved.[\[174\]](#footnote-173) While we often want to resolve tensions by picking one side over another (either/or thinking), paradoxes require us to work with both sides simultaneously (both/and thinking).[\[175\]](#footnote-174) For example, in R\&D we must work with opposing forces: balancing speed with rigor[\[176\]](#footnote-175) (Figure 71), exploring new territories while delivering concrete results, or allowing emergence and serendipity while maintaining focus. Paradoxes are at the heart of complexity,[\[177\]](#footnote-176) challenging us to balance two opposites that cannot be unified, while situations of uncertainty often demand clarity and a clear direction. Working with these tensions is critical to navigating uncertainty.[\[178\]](#footnote-177) ### Participatory Participatory means actively involving people who are often excluded or overlooked in the planning, implementation and evaluation of initiatives that affect their lives. This approach shifts power dynamics[\[179\]](#footnote-178) by creating safe spaces where everyone can contribute equally, while expanding our pool of knowledge by tapping into people's diverse experiences, insights and capabilities. Working this way creates ownership of solutions, as people move from being passive recipients to active shapers of change who understand how their input influences decisions.[\[180\]](#footnote-179) ### Pattern A pattern is a recurring set of signals, relationships, behaviors, or events in a system that reveals how its different parts interact and influence each other over time. Being able to see patterns helps us move beyond individual cases to see the bigger picture, and understand not just what's happening, but why it's happening. While patterns often emerge in local systems, we can also spot them across different contexts and scales. From a global perspective, we look at patterns across our network, where multiple Labs across different contexts are working on similar emerging or urgent issues; whether it's new approaches to waste management, innovative responses to climate change, or evolving digital needs. These patterns serve as indicators that help us shape our global R\&D agenda by revealing where new demands and opportunities are surfacing across diverse contexts. ### People centred People centered is an approach that puts communities and groups at the heart of innovation and design, engaging them deeply to understand and address their challenges through co-creation.[\[181\]](#footnote-180) It is one of our principles. This approach recognizes that communities often understand their challenges best and have valuable insights to contribute, making them active partners in co-creating solutions. Working in a people-centered way means moving beyond quick assessments or assumptions about what people need. Instead, it focuses on building trust, fostering collaboration, and developing solutions that communities can own, maintain, and adapt. This approach tends to favor small, simple interventions over large-scale projects, as these are often more sustainable and better matched to local capabilities. While these interventions may be small in scale, they generate rich insights about the issue at hand, revealing the conditions, capabilities, and resources needed for effective solutions that can be transferred to other contexts. Note that people-centered shares similarities with human-centered approaches, yet specifically focuses on distinct communities rather than universal human needs.[\[182\]](#footnote-181) ### Portfolio A portfolio is a set of interconnected interventions designed and dynamically managed to address complex development challenges.[\[183\]](#footnote-182) A portfolio is more than a collection of thematically related projects. Through portfolios we understand issues from a systems perspective and are able to leverage linkages across interventions to achieve broader transformational outcomes. Think of a portfolio as a learning tool; it brings different stakeholders together to make sense of what they are doing, how it's connected, develop new strategic options, design interventions together, and learn from the effects these interventions produce. As such a portfolio serves as a platform for creating movements of change.[\[184\]](#footnote-183) Accelerator Labs can initiate portfolios as a result of their R\&D activities and are often involved in enabling and facilitating the design and management of portfolios. UNDP has been experimenting with this approach to make it fit for sustainable development, and the practice has evolved and matured over the years. ### Positive deviance Positive deviance looks for individuals or groups who achieve significantly better outcomes than their peers.[\[185\]](#footnote-184) These positive deviants succeed through unique behaviors and strategies, despite facing the same challenges and resources as others.[\[186\]](#footnote-185) To identify and learn from these successful outliers, data plays an important role. We use both traditional data (e.g. official statistics, surveys, interviews) and non-traditional digital data (e.g. mobile records, social media, satellite imagery).[\[187\]](#footnote-186) Especially the rise of digital data and analytics has expanded our ability to spot positive deviants across wider geographies. Finding these actors and learning about their successful practices allows us to leverage their local know-how and share it beyond their communities. ### Power Power[\[188\]](#footnote-187) is the ability to influence outcomes and shape the direction of change within systems, communities and collectives. Power manifests through roles and relationships: who gets to define problems, who is considered an expert, whose reality counts, who makes decisions, and who benefits from solutions.[\[189\]](#footnote-188) Power can be visible, hidden, or invisible.[\[190\]](#footnote-189) Power dynamics exist at every scale: between practitioners and communities, within teams, across organizations, and throughout ecosystems.[\[191\]](#footnote-190) Understanding these dynamics is crucial for R\&D because power shapes the systems we engage with; it can lock systems in place but shifts in power can also trigger transformations and give rise to new systems.[\[192\]](#footnote-191) Exploring and mapping these dynamics ensures that diverse voices are included and positive effects are felt by those who typically have less power.[\[193\]](#footnote-192) ### Practice A practice, in general, refers to a systematic way of working that combines expertise, capabilities, principles, tools and methods to achieve specific outcomes. A practice is often developed, cultivated and employed by a specific group of people who share a common mindset and ways of working. More specifically, for our approach to R\&D, we have identified twelve practices that represent our most crucial "jobs to be done" – activities that move our R\&D process a big step forward. These practices can be employed flexibly in different configurations throughout the entire R\&D journey. ### Principle Principles define the logic of how we do R\&D. They provide coherence and structure to our perceptions (how we see and understand the world), thinking (how we analyze and conceptualize the world) and actions (how we try to transform the world). They involve core beliefs that form our intentions, how we relate to others and how we can push sustainable development forward. They serve as overarching guidelines that should help decision making, create alignment among key ecosystem actors, and prioritize collective learning to shape action. ### Prototyping A prototype is the visible, tangible or functional manifestation of an idea, which we test with others and learn from at an early stage of the development process.[\[194\]](#footnote-193) Prototyping helps us to iterate and improve our ideas and they may also serve as a tool to engage with stakeholders to develop a shared vision or common ground for a solution. ### Question Questions are the starting point for research and learning. From the beginning, since we launched the Accelerator Lab Network, Labs have been developing learning questions for their learning plans. We also develop network-wide learning questions to guide our global R\&D agendas.[\[195\]](#footnote-194) These learning questions define our learning intent and coordinate curiosity among actors, directing our process of exploration, experimentation, and reflection. They enable collective learning[\[196\]](#footnote-195) and help us navigate complexity and uncertainty by focusing our attention on where we need to accelerate our learning. Learning questions aren't set in stone; they improve as we unlock more knowledge. Good questions challenge our assumptions and generate new perspectives, leading to insights and actionable knowledge that help us move forward. ### ​​Reflection Reflection is thinking about past and current experiences to make sense of them and inform future decisions.[\[197\]](#footnote-196) Reflection is a critical part of learning from our actions, observations and experiences,[\[198\]](#footnote-197) it can help us advance our knowledge of the world, change how we think,[\[199\]](#footnote-198) or improve our capabilities.[\[200\]](#footnote-199) "We speed up to act, and slow down to reflect" is our adage. While we naturally reflect on experiences all the time, doing it intentionally and regularly – together or alone – amplifies our learning and improves outcomes. To be meaningful, we belief reflection needs dedicated time and psychological safety,[\[201\]](#footnote-200) especially in groups. Creating spaces and rituals for reflection, for example weekly team check-outs, helps us identify our biases, challenge our assumptions, develop new perspectives, and learn from both success and failure while navigating complexity. ### Reframing Reframing is a way to look at situations or issues from different perspectives, helping us understand them in new ways and discover alternative opportunities.[\[202\]](#footnote-201) When we reframe a situation, we use a different cognitive lens that highlights certain aspects while obscuring others; this selective focus shapes our attention.[\[203\]](#footnote-202) Particularly when working with complex systems, reframing helps us see new options for learning,[\[204\]](#footnote-203) creating space for serendipity[\[205\]](#footnote-204) and pathways for action. The way we communicate these new perspectives[\[206\]](#footnote-205) can be powerful to help other stakeholders understand a situation in a new way and mobilize action. ### Relationship A relationship is a connection between actors that enables information flow, learning, and collective action. In our work we look at relationships in two ways.[\[207\]](#footnote-206) One is through the lens of collective intelligence, the other is through the lens of systems and complex issues. Although there is overlap between the two, the focus of action and intervention can be different. For collective intelligence, relationships form – what Sam Rye calls – the relational infrastructure[\[208\]](#footnote-207) of a community, collective or network. For becoming smarter together and achieving complex goals collectively, understanding how relationships shape an ecosystem and enable information to flow is important. For systems on the other hand we focus more on relational patterns[\[209\]](#footnote-208) that produce certain results, and how we may develop new relational patterns[\[210\]](#footnote-209) (i.e. configurations) for more desirable outcomes. When working with collective intelligence and system transformation, relationships serve as both the enabler and the subject of intervention. It's therefore important to identify which relations are critical for a collective and ecosystem. Because what we don't see, we might inadvertently break.[\[211\]](#footnote-210) In our work, we prioritize activities that build trust and deepen relationships. These meaningful connections increase the space for serendipity,[\[212\]](#footnote-211) build resilience, and enable rapid collective action when opportunities or challenges arise. ### Research & Development Research and Development (R\&D) refers to early-stage innovations and socialized learning[\[213\]](#footnote-212) to close the gap between current results and bigger aspirations, as well as a proven approach to navigate uncharted territory.[\[214\]](#footnote-213) It helps us position ourselves at the edge of sustainable development, developing new knowledge, creating new value and finding entry points for system transformation. While R\&D takes different forms across sectors[\[215\]](#footnote-214) (from advancing scientific knowledge to creating new products in companies to improving public policies or services in government), our approach combines elements of all three. What makes it distinctive is our commitment to collective learning and action at the very core: we actively engage the broader ecosystem – from grassroots to institutional levels and back – enabling its collective intelligence to create collective impact.[\[216\]](#footnote-215) Within these ecosystems, we mobilize and strengthen existing collectives, or where necessary catalyze new ones. We help them see the potential and benefits of addressing sustainable development challenges together, and make use of the synergies across their efforts. ### Scaling Scaling typically refers to an increase in size, quantity or reach. It is often assumed that scaling occurs through linear growth, through replication of successful solutions to different locations, with the expectation that results will grow in direct proportion to the resources invested. However, in sustainable development scaling is much more multifaceted and unpredictable, presenting many possible trajectories and timelines. Different aspects of our work scale through different logics with various aims. Solutions, knowledge, relationships, power, practices, conditions, outcomes, and effects[\[217\]](#footnote-216) may scale by: reaching more people and places through diffusion, influencing higher-level policies and institutions, or impacting cultural norms and beliefs as powerful levers of change.[\[218\]](#footnote-217) While these diverse scaling pathways can facilitate systemic change, our experience shows that transformation rarely stems from a single, big idea. Instead, system change often arises from the deliberate accumulation of many, often small, solutions and interventions.[\[219\]](#footnote-218) It's worth remembering that not every innovation will, or needs to, scale.[\[220\]](#footnote-219) Scaling fundamentally occurs through the ecosystem.[\[221\]](#footnote-220) Ecosystems help cultivate, connect and legitimize collections of solutions. This typically happens through network effects, such as when the right actors connect, when solutions emerge to shed light on unmet needs, or when new technologies create new markets or solve stubborn public sector problems. ### SDG SDGs (Sustainable Development Goals) are 17 interconnected global goals established by the United Nations to achieve a better and more sustainable future for all by 2030.[\[222\]](#footnote-221) They serve as a shared blueprint that helps governments, organizations, and communities work together toward common aims like ending poverty, improving health and education, reducing inequality, tackling climate change, and preserving our environment. The SDGs provide a common language and framework for measuring progress, while encouraging partnerships and innovation to overcome persistent development challenges. ### Sensing Sensing is an activity within our R\&D modes that focuses on understanding what's going on in the ecosystem and identifying where acceleration is needed. Through sensing, we look for the new. It involves continuously scanning the environment for weak signals of change, emerging challenges, and unaddressed needs that may not be on our radar yet. ### Serendipity Serendipity is the act of making an accidental, yet surprising and valuable, discovery.[\[223\]](#footnote-222) We can't control or predict it, and it won't happen on demand – but we can create conditions that make it more likely to happen.[\[224\]](#footnote-223) In R\&D, this means deliberately expanding our opportunity space for discoveries by staying open to unexpected findings and being ready to act on them when they emerge. It's about unlocking adjacent possibilities, seeing the dots and their connections that others don't. As Louis Pasteur reminds us: "In the field of observation, chance favors only the prepared mind.” ### Signal Signals are data points that indicate a potentially significant change which is about to happen.[\[225\]](#footnote-224) As William Gibson famously observed: "The future is already here, it's just not evenly distributed." Signals in that sense point us to pockets where the future is already happening. They often appear as surprises or anomalies at the periphery of our attention in unexpected places like new community practices, unusual policies, or technological experiments. While these signals might seem small or easy to dismiss at first, they can point to important systemic changes ahead, particularly when there is an increase in their strength, frequency, spread or maturity[\[226\]](#footnote-225). By actively "scanning the horizon" and looking for signals, capturing and documenting them, we can identify patterns of possible futures emerging.[\[227\]](#footnote-226) This helps us make sense of uncertainty, anticipate what's coming and identify both risks and opportunities early on. ### Skill Skills are specific abilities that we use to perform R\&D tasks and achieve results. Skills can develop through training or learning by doing, but not all skills are teachable, particularly complex skills[\[228\]](#footnote-227) usually develop through experience. In our R\&D work we broadly use two types of skills, soft and technical skills. Soft skills are interpersonal abilities that help us create conditions for change, like building relationships, getting buy-in from decision makers, and helping stakeholders see things differently. Technical skills are practical abilities we use to learn about problems, identify solutions, and test ideas, like data analysis, prototyping, or using digital tools. Both types of skills often work in tandem in R\&D. For instance, when running an experiment, we need technical skills to design it properly, and soft skills to engage stakeholders effectively. Understanding what skills are required for R\&D helps us design effective teams and guides our approach to recruitment and training of R\&D talent. ### Solutions mapping Solutions mapping is a method to discover innovations that already exist within an ecosystem, document them, and make this knowledge accessible to others. It is one of the principal methodologies of the Accelerator Labs.[\[229\]](#footnote-228) When mapping solutions, we look for people solving problems and putting bandaids on broken systems: grassroots innovators, positive deviants, user-innovators[\[230\]](#footnote-229) and start-ups. We particularly focus on grassroots innovators as they give us deep insight into how the people most affected solve their problems and help us learn about their deep local knowledge. By mapping solutions, we deliberately explore the adjacent possible which increases our chances of serendipitously finding innovations that can help our R\&D process take significant steps forward. ### Space A space is a cognitive, conceptual or collective construct that enables us to coordinate our learning and understanding, focus our attention, set the scope of our work, or create the enabling conditions for action. We often use the term “space” fluidly and metaphorically; it can refer to an innovation or R\&D space where we have a mandate to explore and experiment, a learning space where people feel safe to share and reflect on their experiences, an information space that holds insights about issues or visions of potential solutions, or a social space that gives identity to a group and enables collaboration. While we primarily talk about conceptual spaces here, they often connect with physical spaces. For example, the physical setup of an innovation lab can promote creativity, sense making and experimentation.[\[231\]](#footnote-230) Creating and nurturing conceptual spaces is crucial for R\&D, as they provide structure to learning, reflecting, deciding, and acting, while holding the intent of a collective. ### Sustainable development The principles of sustainable development are anchored in the 2030 Agenda,[\[232\]](#footnote-231) Sustainable Development Goals (SDGs), adopted by all UN Member States in 2015 as a universal call to action with specific targets to end poverty, protect the planet, and ensure prosperity for all by 2030. Sustainable development aims to meet the needs of the present without compromising the ability of future generations to meet their own needs.[\[233\]](#footnote-232) This concept balances three interconnected dimensions: economic growth, social inclusion, and environmental protection. Rather than treating these as separate challenges, sustainable development recognizes they must be tackled together and in partnership with communities, businesses and governments. This aspect makes sustainable development an area that needs R\&D, no country has achieved a full environmental, economic and social balance, so experimentation is needed to find out what works. ### System A system is a set of elements connected together in a way that produces its own pattern of behavior over time.[\[234\]](#footnote-233) When we talk about systems in development, we encounter them everywhere, from local food systems with their farmers, markets and supply chains, to global systems like climate and finance. We focus especially on complex systems where many actors and factors interact in unpredictable ways[\[235\]](#footnote-234): communities adapting to climate change, or economies transitioning to clean energy. Understanding systems helps us see past immediate symptoms to uncover the underlying conditions and understand the deeper connections[\[236\]](#footnote-235) and patterns that create and sustain problems. Transforming these systems is a dynamic process;[\[237\]](#footnote-236) it requires continuous learning and adaptation. Through R\&D, we explore these systems, map their connections, find the cracks and work with them,[\[238\]](#footnote-237) test potential solutions, and learn what works to achieve better, more equitable outcomes. ### Tactic Tactics are the specific "moves" or actions we take to create momentum for R\&D, gain support and legitimacy, inform decisions, and get things moving. Tactics are the tricks of the trade that help us create the conditions for R\&D.[\[239\]](#footnote-238) This guide was written based on collecting tactics across labs in 115 countries over 5 years. These tactics are integrated throughout the guide, with some explicitly named and others embedded in the practices, methods, and principles we describe. ### Technology Technology refers to the tools and systems we create and use to solve problems and achieve our goals. It comes in two main forms: physical technologies like tools, machines, devices and materials (from simple sticky notes to complex computers), and digital technologies like software, algorithms and online platforms. These technologies augment human capabilities in countless ways, from sensors that extend our perception, to AI that enhances our ability to analyse large amounts of data and information, to networks that enable global collaboration and instant communication. But technology alone isn't enough; it needs to be combined with methods to be effective. In R\&D, we experiment with new configurations of technology,[\[240\]](#footnote-239) data and people to create collective intelligence by leveraging the wisdom of crowds, enabling collective learning, supporting decision making, and coordinating collective action. This allows us to do R\&D at different scales, from hyper local to global, enabling us to share insights and diffuse innovations across communities and regions at lightning speed.[\[241\]](#footnote-240) While technology can help us include voices or reach communities that are often left out, those lacking access or digital skills are excluded from these opportunities, which may deepen inequalities over time. ### Testing Testing is an activity within our R\&D modes that focuses on finding out what works and probing how a system responds. Through testing, we learn how our interventions perform in real conditions, identify assumptions, improve our ideas, and create evidence to mobilise action, learning and resources. ### Tool Tools are the instruments that enable R\&D activities, typically linked to specific methods or purposes.[\[242\]](#footnote-241) Tools provide structure to an R\&D process. They come in many forms, with varying levels of complexity and accessibility. Easily accessible tools like worksheets and canvasses[\[243\]](#footnote-242) help structure conversations, support reflection, and identify our biases and assumptions. They also assist in generating insights, developing solutions, mapping system dynamics, and coordinating action. An often overlooked but powerful category includes tangible artifacts like Lego Serious Play, which helps groups explore and map systems while expressing mental models, as "embodied metaphors", to build shared understanding.[\[244\]](#footnote-243) At the other end of the spectrum are sophisticated technologies such as AI systems, natural language processing, drones, sensors, and data analytics platforms, all designed for collecting, analyzing, and presenting data. While tools can be used "off the shelf," they are often adapted to specific needs or developed from scratch to meet unique requirements. ### Uncertainty Uncertainty is a situation of not-knowing.[\[245\]](#footnote-244) We consider a situation to be uncertain when we have incomplete or no knowledge of what is happening, how an intervention might perform, what actions we can take, or what outcomes might emerge. While uncertainty and risk are often used interchangeably in everyday language, they are distinct concepts. Risk can be calculated based on probabilities using existing data (e.g., how likely a solution is to succeed), whereas uncertainty exists when we lack the data or prior knowledge to calculate such probabilities.[\[246\]](#footnote-245) Our R\&D practice focuses on navigating new spaces in sustainable development with conditions of high uncertainty, and finding pathways to positive change. ### Value Value describes the positive changes, benefits, or improvements that R\&D activities create for people, communities, organizations, or the environment. When doing R\&D, we create value in multiple ways that reinforce each other: from amplifying citizens' voices and democratizing development, to detecting emerging issues, building collective intelligence in ecosystems, de-risking innovation through experimentation, using data and digital solutions, and creating compelling narratives that inspire action.[\[247\]](#footnote-246) This goes beyond just monetary worth to include social impact, environmental sustainability, increased knowledge, stronger relationships, and improved capabilities. Sometimes value is immediate and obvious, like a new service that helps people. Other times it's indirect or takes time to emerge, like when experimentation leads to insights that transform how ecosystems work. In ecosystems, value is interconnected and flows between actors through their interactions. What creates a benefit for one actor might create a disadvantage for another, which means we need to think beyond a single value chain and understand the value dynamics as networked.[\[248\]](#footnote-247) ### Working out loud Working Out Loud is a communication strategy that involves documenting and sharing insights, updates and reflections openly as it happens. Unlike traditional reporting that happens at predefined project milestones, Working Out Loud happens throughout the journey, with frequency varying from weekly to monthly or perhaps less. We share the full spectrum of our work – from successes to failed experiments – and what we learn when things take unexpected turns. Frequently documenting and sharing updates helps us reflect on our experiences, capture insights, build a knowledge repository, advance our practice over time, and enables others to learn from our work. We usually work out loud through blog posts, podcasts, or vlogs, choosing the format that best suits our audience. Consider working out loud as an engagement instrument. By telling both our ecosystem and the wider world what we're doing, we create opportunities for attracting potential collaborators and making our work visible to unusual suspects. This openness creates opportunities for serendipity, enabling new connections between people, insights, and ideas. Time and again, ecosystem actors have reached out to partner with us after discovering our work through our blogs and social media. *** ## **Notes** 1. Based on Loreto et al.(2016) [↑](#footnote-ref-0) 2. Johnson (2010a, p. 31) [↑](#footnote-ref-1) 3. Stuart Kauffman (1995) introduced the concept of the adjacent possible in his works on evolutionary biology and complex systems. For an engaging introduction to how it drives innovation, see Chapter 1 (pp. 25-42) of Steven Johnson's Where Good Ideas Come From (2010a). Björneborn (2023) provides a more academic introduction, discussing the state of the art and different types of adjacent possibles. [↑](#footnote-ref-2) 4. Johnson (2010a, p. 31) [↑](#footnote-ref-3) 5. See the paper “Curiosity and networks of possibility” by Perry Zurn & Dani Bassett (2023) for how the adjacent possible and curiosity are connected. [↑](#footnote-ref-4) 6. Note, Agency differs from power in important ways. While agency is the capacity to act and make decisions, power shapes whether those actions can create real change in systems. Actors might have agency (the ability to make choices) but lack power (the ability to make those choices matter). Conversely, those with power (influence over outcomes) might choose not to exercise their agency. See also: Power. [↑](#footnote-ref-5) 7. For a practical and accessible introduction to ambiguity, see "Navigating Ambiguity" by Andrea Small and Kelly Schmutte (2019). [↑](#footnote-ref-6) 8. Crawford (2024) [↑](#footnote-ref-7) 9. See van Es et al. (2015, pp. 20-25); in their guide, they explain the role of assumptions in developing theories of change. [↑](#footnote-ref-8) 10. Vogel (2012) [↑](#footnote-ref-9) 11. Kahneman (2011) [↑](#footnote-ref-10) 12. See Liedtka (2015) [↑](#footnote-ref-11) 13. For Lucarelli (2023b) for a brief explanation how we curate patterns for bottom-up organizational learning. For examples, see our report on informal or popular transportation (Nebrija et al., 2024), or blogs on how digitalization is changing what informality looks like (Akinyemi, Leurs & Lucarelli, 2022) and how communities are adapting to climate change (Cottica, 2023). [↑](#footnote-ref-12) 14. Levi-Strauss (1966) [↑](#footnote-ref-13) 15. Mateus & Sarkar (2024) [↑](#footnote-ref-14) 16. When people solve problems in everyday life, they often discover what they need at the same time as they find a solution, testing them together as a "need-solution pair" Von Hippel & Von Krogh, 2016). [↑](#footnote-ref-15) 17. See Stock-Homburg, Heald, Holthaus, Gillert & Von Hippel (2021). [↑](#footnote-ref-16) 18. Lennart Björneborn (2017) refers to these possibilities as “usage potential”, the affordances that facilitate serendipity. [↑](#footnote-ref-17) 19. Johnson (2010a, p. 41) [↑](#footnote-ref-18) 20. Based on Stewart Brand's The Clock of the Long Now (1999, p. 37). [↑](#footnote-ref-19) 21. Brand (2018), or see how NOBL (2021) translated this idea to an organizational context. [↑](#footnote-ref-20) 22. Also see the chapter on time and scales in the “The art of systems change” by the Fuller Transformation Collaborative (2019, pp. 57-63). [↑](#footnote-ref-21) 23. The differences between co-creation (e.g., Gouillart & Hallett, 2015), co-design (e.g., McKercher, 2020), co-production (e.g., SCIE, 2022), participatory design (e.g., Interaction Design Foundation, 2023), and other participatory approaches can be subtle, and these terms are sometimes used interchangeably. While a detailed exploration of these differences falls outside this guide's scope, we refer to Vargas, Whelan, Brimblecombe, and Allendera (2023) for an in-depth discussion. [↑](#footnote-ref-22) 24. Pfortmüller (2022); Mintzberg (2015) [↑](#footnote-ref-23) 25. See Nesta's (2019) playbook for an introduction and hands-on guidance to design collective intelligence. For a comprehensive overview of the field, including its background and fundamental principles, we recommend Geoff Mulgan's “Big Mind” (2018). [↑](#footnote-ref-24) 26. We have published a collection of case studies (see Berditchevskaia, Peach, Lucarelli & Ebelshaeuser, 2021) illustrating how collective intelligence is being used by the UNDP Accelerator Labs for sustainable development. For a deeper analysis, see our accompanying report (Peach, Berditchevskaia, Mulgan, Lucarelli & Ebelshaeuser, 2021), which maps the collective intelligence approaches most frequently used, and often combined, to address sustainable development challenges. In addition, see our UNTAPPED report (Berditchevskaia, Albert, Peach, Lucarelli, & Cottica, 2024), showcasing a collection of case studies on how the Accelerator Labs use collective intelligence for climate action. [↑](#footnote-ref-25) 27. Johnson (2010b) [↑](#footnote-ref-26) 28. See Arthur (2009, p. 21) [↑](#footnote-ref-27) 29. Every innovator is an odd ball according to Anil Gupta (2016, p. 14). They stand out and are non-conventional in their thinking and finding ways to make their innovations work. In particular, grassroots innovators persevere to improve their innovations, even when faced with significant financial and social challenges and long struggles. [↑](#footnote-ref-28) 30. Hess & Ostrom (2007) [↑](#footnote-ref-29) 31. See Lucarelli (2023b), or visit our SDG Innovation Commons platform: [↑](#footnote-ref-30) 32. Potts, Torrance, Harhoff, & Von Hippel (2021) [↑](#footnote-ref-31) 33. See Potts (2019, p. 7); also see his comments on innovation problems essentially being knowledge problems, which require distributed specialized knowledge to be recombined in order to discover new opportunities, knowledge and sources of value (Potts, 2019, pp. 47-48). [↑](#footnote-ref-32) 34. Potts (2019, p. 1) [↑](#footnote-ref-33) 35. Lucarelli (2023b) [↑](#footnote-ref-34) 36. "The essence of community is relationships of care," according to Erin Dixon and colleagues (2024) from Community Weaving. For practical guidance on building and curating communities, we highly recommend their guide "Community Weaving: A Framework for Weaving Healthy Communities" (Dixon et al., 2024). Also recommended for those building communities in an institutional context is the guide "Cultivating UN Innovation Communities: Key Questions and Considerations" (UN Innovation Network, 2025). [↑](#footnote-ref-35) 37. Pfortmüller (2022) and Mintzberg (2015) [↑](#footnote-ref-36) 38. See DS4SI (2020) [↑](#footnote-ref-37) 39. Winhall & Leadbeater (2022) [↑](#footnote-ref-38) 40. See Snowden (2015) [↑](#footnote-ref-39) 41. See Hofstede & Hofstede (2005) [↑](#footnote-ref-40) 42. See Wolf (1958) [↑](#footnote-ref-41) 43. Cosmovisions are comprehensive worldviews that explain the relationships between humans, nature, the spiritual realm, and the cosmos. [↑](#footnote-ref-42) 44. Hall (1983); Hall (1990) [↑](#footnote-ref-43) 45. For more on the art and craft of curating networks and collectives, we recommend David Ehrlichman's "Impact Networks" (2021). Note that Ehrlichman refers to curation as "cultivation" (p. 57), drawing an analogy with gardens – networks, like gardens, need to be cultivated in order to grow and flourish. We however prefer "curation" as a term because it captures the intentional, purpose-driven nature of our work to progress collectively on the sustainable development goals. [↑](#footnote-ref-44) 46. Zurn & Bassett (2022, p. 5); Lydon-Staley et al. (2021). [↑](#footnote-ref-45) 47. Zurn & Bassett (2023) [↑](#footnote-ref-46) 48. Zurn & Bassett (2023) [↑](#footnote-ref-47) 49. In their book, Zurn and Bassett (2022, pp. 98-107) use metaphorical labels for these approaches: the hunter (goal-oriented), the dancer (associative), and the busybody (serendipitous). We've chosen not to adopt these metaphors in this description but use descriptive terms instead to make the strategies immediately clear. Also see Zurn (2019). [↑](#footnote-ref-48) 50. Kelleher & Tierney (2018) [↑](#footnote-ref-49) 51. For a more comprehensive overview of non-traditional data sources, see “A Guide to Data Innovation for Development” (UNDP & UN Global Pulse, 2017, p. 19) or Nesta's “Collective Intelligence Design Playbook” (Peach et al., 2020, p. 92). Also see the overview of citizen science initiatives in Argentina (National Ministry of Science, Technology and Innovation, & UNDP Argentina,2023) [↑](#footnote-ref-50) 52. UNDP & UN Global Pulse (2017, p. 18) [↑](#footnote-ref-51) 53. One of the principles for designing collective intelligence is “data empowerment, not data extraction.” (Peach et al., 2020, p. 40). [↑](#footnote-ref-52) 54. In practice, integrating multiple knowledge sources can present challenges, particularly when these sources are grounded in different epistemologies; we often refer to this as the “battle of epistemologies.” What counts as valid knowledge and how we justify it varies across scientific, traditional, local, and other knowledge systems (see Nakashima & Elias, 2002; Gregory, 2025; Yunkaporta, 2019). We need to create spaces where these different ways of knowing can coexist and complement each other, rather than compete or invalidate one another. Creating synergies between knowledge systems helps us generate a more comprehensive understanding of development challenges and opportunities. [↑](#footnote-ref-53) 55. For examples see our report “Next practices for a more sustainable future“ (UNDP Accelerator Labs, 2025) or “Modernizing development: Introducing portfolios” (UNDP, 2025\_)\_ [↑](#footnote-ref-54) 56. In his book "Diffusion of Innovations", Everett Rogers (2003) provides a comprehensive introduction to diffusion theory, explaining how innovations spread through social systems. [↑](#footnote-ref-55) 57. These three terms correspond to progressive levels of absorption and ownership: adopt (using as-is), adapt (modifying to fit context), and enhance (adding functions or improving functionality). Also see Nippard, Hitchins & Elliott (2014) who use a similar logic in their Adopt-Adapt-Expand-Respond framework for measuring systemic change. [↑](#footnote-ref-56) 58. See Ang (2018a) and Ang (2018b). [↑](#footnote-ref-57) 59. See Lucarelli (2019) for a short explanation of the concept of directed improvisation for the onboarding of the Labs. [↑](#footnote-ref-58) 60. See Linda Doyle's white paper on the "Vector Theory of Change" (2021). [↑](#footnote-ref-59) 61. For a critical analysis of “ecosystem” as a metaphor and a comparative examination of biological versus organizational ecosystems, see Mars, Bronstein, and Lusch (2012). [↑](#footnote-ref-60) 62. See James Moore's "Predators and Prey" (1993), a classic work on using ecosystem approaches as business strategy. For insights on how ecosystem approaches can create value across multiple actor (i.e. businesses, non-profit organizations, customers, and society) see Elke den Ouden's "Innovation Design" (2012). [↑](#footnote-ref-61) 63. MIT D-Lab has developed an excellent framework for analysing and understanding of place-based innovation ecosystems, see Hoffecker (2019). For an example of what ecosystem mapping looks like, we recommend reviewing the detailed mapping of Kenya's innovation ecosystem conducted by UNDP Accelerator Lab Kenya (see Kiarie-Kimondo, et al., 2022). Also worth reading is Liz Altman and Frank Nagle's (2022) analysis of the Accelerator Labs, where they describe the network as an interconnected system of local innovation ecosystems and highlight valuable lessons the private sector can learn from this initiative. [↑](#footnote-ref-62) 64. That also means, there is no such thing as (eco)system leadership (see Needham, Gale and Waring, 2025). [↑](#footnote-ref-63) 65. See Fuller, Jacobides, & Reeves (2019) [↑](#footnote-ref-64) 66. In ecosystems, relationships appear to be either competitive or collaborative, but the reality is more nuanced. Actors must both compete and collaborate (Moore, 1993; Fuller et al. 2019) – forming symbiotic relationships (see Yoon, Moon & Lee,2022, Good Shifts, 2024) – to position themselves and create unique value propositions or impact in a multi-actor business environment. [↑](#footnote-ref-65) 67. Cohendet, Grandadam, & Simon (2010) [↑](#footnote-ref-66) 68. Cohendet, Grandadam, & Suire (2021) [↑](#footnote-ref-67) 69. See Christian Bason’s lates manifesto on ecosystem transitions (Bason, 2025). [↑](#footnote-ref-68) 70. Pfortmüller (2022); Mintzberg (2015) [↑](#footnote-ref-69) 71. DS4SI (2020) [↑](#footnote-ref-70) 72. Lucarelli (2019) [↑](#footnote-ref-71) 73. Wheatley and Frieze (2006) explain how emergence helps scale social innovation. [↑](#footnote-ref-72) 74. Traditional development planning relies heavily on results-based management tools like logical frameworks (logframes), which assume linear cause-and-effect relationships: if we do A, it leads to B, then C happens. As Gina Lucarelli (2019) suggests, we might instead think of theories of change as radial rather than linear – recognizing multiple possible pathways to favorable outcomes that cannot always be predicted in advance. [↑](#footnote-ref-73) 75. Kouprie and Sleeswijk Visser (2009) provide a comprehensive introduction to empathy, its role in design, and the process behind it. [↑](#footnote-ref-74) 76. Breckon (2016). [↑](#footnote-ref-75) 77. Christensen, Leurs & Quaggiotto (2017) explain that experimentation can best be seen as a continuum of different approaches rather than as one method, from probing to trial and error and controlled trials, with different methods used depending on whether solutions and their intended outcomes are known, partially known, or not known at all. Building on this, a group of Heads of Experimentation across the UNDP Accelerator Labs (see Pop Ivanov et al., 2025) have reflected on their practice and identified four essential components of any experiment (question, hypothesis, data, and analysis) where the degree of control over each element determines the level of rigor that's feasible. This ranges from adaptive experiments with limited control to highly controlled trials. [↑](#footnote-ref-76) 78. UNDP Accelerator Labs (2023) [↑](#footnote-ref-77) 79. Edmondson (2011). [↑](#footnote-ref-78) 80. Leurs & Roberts (2018, p. 72-73) [↑](#footnote-ref-79) 81. Birney (2021b) [↑](#footnote-ref-80) 82. Brown (2021) [↑](#footnote-ref-81) 83. See Barry’s (2021) [↑](#footnote-ref-82) 84. For more on psychological safety, see Amy Edmondson's work, particularly her book "The Fearless Organization" (Edmondson, 2018) for practical guidance, or her foundational paper with Lei (Edmondson & Lei, 2014) for theoretical understanding. [↑](#footnote-ref-83) 85. Fernández Sirera et al. (2025) provide insightful guidance on navigating polarities when building shared agendas among diverse stakeholders. [↑](#footnote-ref-84) 86. Edmondson (2011) [↑](#footnote-ref-85) 87. See Snowden (2012) [↑](#footnote-ref-86) 88. See Wang (2016) for an overview of formats to celebrate failure, e.g. Fail Fests and F\*\&k Up Nights. [↑](#footnote-ref-87) 89. See for example Dave Gray’s growing collection of visual frameworks on [www.visualframeworks.com](http://www.visualframeworks.com). [↑](#footnote-ref-88) 90. See Chipchase (2024) [↑](#footnote-ref-89) 91. See for example Hugh Dubberly’s (2004) compendium of design process, or VanPatter and Pastor’s (2018) collection of innovation methods. [↑](#footnote-ref-90) 92. See for example Nesta’s Playbook for Innovation Learning (Leurs & Roberts, 2018). [↑](#footnote-ref-91) 93. As such, frameworks function as "boundary objects" when they enable actors from different social worlds to cross boundaries (see Akkerman & Baker, 2011) interact with each other. In order to do that, these frameworks need to be plastic enough to adapt to individual realities and local needs, yet robust enough to maintain a common identity across parties (Star & Griesemer, 1989; Star, 2010). [↑](#footnote-ref-92) 94. Effective memes are memorable (Heath & Heath, 2007) and propagate themselves by spreading from one mind to the other (Dawkins, 1989). [↑](#footnote-ref-93) 95. Radjou & Prabhu (2015); Radjou, Prabhu & Ahuja (2012, pp. 64-82). Also see Anil Gupta’s (2016, pp.242-300) reflections on the culture of frugality. [↑](#footnote-ref-94) 96. Gustilo Ong & Gustale (2020), also see Radjou, Prabhu & Ahuja (2012) on Jugaad Innovation. However, as Anil Gupta notes (2016, p. 265), we need to bear in mind that the Jugaad mindset generally focuses on getting around problems rather than actually solving them. Jugaad innovators don't plan; they improvise (also see Radjou et al, 2012, p. 104) and solve problems in an ad hoc fashion, which is acceptable in the short term but can affect the growth of an innovation culture in the long term. [↑](#footnote-ref-95) 97. Gupta (2016, p. 271); Gustilo Ong & Gustale (2020) [↑](#footnote-ref-96) 98. Radjou, Prabhu & Ahuja (2012, pp. 99-111) [↑](#footnote-ref-97) 99. This spectrum is often depicted as a cone, also known as the “futures cone.” See Joseph Voros' (2017) blog for a short history of this concept and a comprehensive overview of the different types of futures to consider. [↑](#footnote-ref-98) 100. As Scott Smith (2020, pp. 46-47) reflects in “How to Future”: those who explore and map future scenarios may not be the ones impacted by them – so “whose future is it anyway?" [↑](#footnote-ref-99) 101. Social imagination is an emerging practice that has been adopted by a number of Labs. See the work of Geoff Mulgan (2020) and Cassie Robinson (2022) for an introduction. [↑](#footnote-ref-100) 102. This diagram builds on the work of Charles Leadbeater (2009), Sabine Junginger (2014). Here, "designers" refers broadly to anyone who shapes solutions: policy makers, R\&D specialists, service or product designers, or development practitioners. Leadbeater observes that organisations often claim to work "for" people while actually doing things "to" them, processing them through systems designed without their input. Junginger similarly notes that designing "on behalf of" someone reflects a paternalistic approach that positions designers as experts who know best. [↑](#footnote-ref-101) 103. From Lucarelli (2023a). For an in-depth exploration of grassroots innovation, we highly recommend Gupta's book “Grassroots Innovation: Minds on the Margin Are Not Marginal Minds” (2016). We also suggest watching “For Tomorrow” (For Tomorrow 2030, 2021), our documentary showcasing these innovations in action. [↑](#footnote-ref-102) 104. Yuen Yuen Ang (2021) points out that common thinking about innovation often suggests that only wealthy, city-dwelling, well-educated, and tech-savvy people are capable of innovation. However, the reality is that the non-elites, those who are poor and have limited resources, innovate all the time; it is essential for their survival. In doing so, they demonstrate an innate ingenuity and a deep understanding of their environment that deserves more attention. This echoes what Anil Gupta has long maintained: "Minds on the margin are not marginal minds" (Gupta, 2016). [↑](#footnote-ref-103) 105. Visit our SDG Innovation Commons () for a curated collection of solutions and grassroots innovations that we can share openly, with the consent of the innovators. [↑](#footnote-ref-104) 106. Grassroots innovation raises questions about who exercises agency in innovation processes and who has the best information to address issues. Traditionally, problems are solved for people and communities. The emergence of co-creation over recent decades has changed this dynamic: people affected by issues are not just recipients or users of solutions but active co-creators (see Sanders & Stappers, 2008, 2012; Leadbeater, 2009; Manzini, 2015). Grassroots innovation goes further, building on the innate capacity of people to solve their own problems. [↑](#footnote-ref-105) 107. Gupta (2012), also see the book “Jugaad Innovation” by Navi Dadkou, Jaideep Prabu and Simone Ahuja (2012). [↑](#footnote-ref-106) 108. Ang (2021) [↑](#footnote-ref-107) 109. See for example Nripen Kalita’s zero-head water turbine: [↑](#footnote-ref-108) 110. See for example Dyonisius et al. (2024, pp. 41-47). [↑](#footnote-ref-109) 111. See for example Bennett (2023) [↑](#footnote-ref-110) 112. Gupta (2016, p. 271); Gustilo Ong & Gustale (2020) [↑](#footnote-ref-111) 113. These align with Eric von Hippel and Georg von Krogh's (2016) concept of “need-solution pairs,” instances where problems and their solutions are discovered simultaneously in real-world contexts. [↑](#footnote-ref-112) 114. Schrage (2014) [↑](#footnote-ref-113) 115. Johnson (2010a) [↑](#footnote-ref-114) 116. Schrage (2014) [↑](#footnote-ref-115) 117. Mulgan (2014) [↑](#footnote-ref-116) 118. Glennie et al. (2020); Klingler-Vidra et al. (2022) [↑](#footnote-ref-117) 119. Quaggiotto (2020) [↑](#footnote-ref-118) 120. Bason (2010), also see Leurs & Roberts (2018, pp. 62-63) [↑](#footnote-ref-119) 121. See Johnson (2010a, p. 35) [↑](#footnote-ref-120) 122. Mulgan (2014) [↑](#footnote-ref-121) 123. See our annual report 2023 (UNDP Accelerator Labs, 2024a). [↑](#footnote-ref-122) 124. In this glossary, we mainly focus on defining "issue," but also refer to similar terms like "(complex) problems" and "(development) challenges." While these words mean slightly different things in some situations, we use them interchangeably to keep things simple. [↑](#footnote-ref-123) 125. The distinction between technical and adaptive problems comes from Ronald Heifetz and Marty Linsky's (2002) work on leadership and organizational change; also see their book on adaptive leadership (Heifetz, Grashow, Linsky, 2009, pp. 19-23). Other useful classifications exist as well. Horst Rittel and Melvin Webber (1973), in their study on social or policy planning, identified two classes of problems: tame (well-defined with clear parameters and known solution paths) and wicked (ill-defined, complex, and constantly evolving) – which apparently shares many similarities with Heifetz and Linsky's categories. Additionally, Dave Snowden's Cynefin framework (see Snowden & Boone, 2007) for decision making is often used to navigate the space of complex problems. He distinguishes between four domains: simple (obvious cause-effect relationships, best practices apply), complicated (cause-effect relationships discoverable through analysis, good practices apply), complex (emergent patterns, unpredictable cause-effect relationships, requires probing and sensing), and chaotic or crisis (no discernible cause-effect relationships, requires immediate action). [↑](#footnote-ref-124) 126. This is essentially a reductionist approach; by breaking down a system into its constituent parts, we understand how each element independently affects the final result. Technical problems are solved by fixing the parts that are broken or not performing well. [↑](#footnote-ref-125) 127. Rittel and Horst (1973) refer to the absence of a clear end-state as if there is “no stopping rule”, as understanding the nature of a problem will always be insufficient and there are no ends to causal chains that are linked to other systems. Also, as many stakeholders are involved, there are no shared criteria when a problem is solved; at best, an improved situation may be “good enough” or “less worse”, and just a temporary state. Problems without stopping roles are basically what James Carse (1986) describes as infinite games. Infinite games have no fixed rules; they change and are renegotiated all the time. There is no end, and there are no clear winners or losers. Finite games, by contrast, are played with the purpose of winning, have fixed rules, and a clear beginning and end with winners and losers. Technical problems often resemble finite games (with clear endpoints), while adaptive challenges function more like infinite games (requiring ongoing adaptation and participation). [↑](#footnote-ref-126) 128. Heifetz, Grashow & Linsky (2009, p. 19) [↑](#footnote-ref-127) 129. Defining a lab can be challenging given the diverse range of lab types and approaches that exist. For a comprehensive discussion of different lab types and their characteristics, see The Future of Labs Primers Report (Action Lab & Social Innovation Canada, 2024). Also see Nesta’s practice guide (Puttick, 2014) for a quick introduction and practical guidance. [↑](#footnote-ref-128) 130. While learning can also refer to the development of skills, in this guide we primarily focus on learning as the acquisition and development of knowledge about the world. [↑](#footnote-ref-129) 131. Lowe et al. (2022) [↑](#footnote-ref-130) 132. Schön (1984) [↑](#footnote-ref-131) 133. Mulgan (2018); Mulgan (2017, pp. 70-75) [↑](#footnote-ref-132) 134. For more on the concept of unlearning, see Mark Bonchek's work "Why the Problem with Learning Is Unlearning" (2016), where he explains that unlearning is not about forgetting, but adopting alternative mental models. [↑](#footnote-ref-133) 135. By expert we don’t mean just subject matter experts with deep conceptual understanding of a policy area, people can also be experts of their own experiences (see Sleeswijk Visser et al., 2005). [↑](#footnote-ref-134) 136. For more background information and practical guidance on learning circles, see our publication “A rough guide for running learning circles” (Leurs, Akinyemi, & Lucarelli, 2025). [↑](#footnote-ref-135) 137. Tacit knowledge is inherently difficult to articulate or convey through written or verbal means (Polanyi 1966/2009, p. 4). But it is rich and contextual. It's knowledge to be discovered – not to be extracted. [↑](#footnote-ref-136) 138. Snowden (2008). [↑](#footnote-ref-137) 139. In management research there is growing interest in the process and practice of legitimacy (see Suddaby, Bitektine & Haack, 2017; Siraz, Teo & Prashantham, 2023). For a more accessible introduction to this topic, see Dave Kiwi's blog at space-for-change.com/change-blog. [↑](#footnote-ref-138) 140. For practical guidance, we recommend looking at "The Legitimacy Lens," a toolkit developed by The Institutional Architecture Lab (TIAL) for institutional entrepreneurs who create, design, lead, and support new or updated institutions (Leppänen, Seddon, & Blake, 2025). This resource offers frameworks and reflection questions for building and renewing institutions. Yet, we see its relevance for establishing innovation and R\&D functions outside an institutional context as well. [↑](#footnote-ref-139) 141. Meadows (1999, p. 1) [↑](#footnote-ref-140) 142. For a detailed description of Donella Meadows's (2008, pp. 145-165) twelve leverage points, see her book Thinking in Systems: A Primer, or see Meadows (1999). [↑](#footnote-ref-141) 143. FAO (2006) [↑](#footnote-ref-142) 144. See for example the work the Accelerator Lab has been doing in Fiji, one of the large ocean states that is increasingly experiencing the effects of climate change (Kakal, 2020). [↑](#footnote-ref-143) 145. The distinctions between these knowledge systems are quite subtle, and discussing them would go beyond the scope of this guide. For a brief introduction, we refer to "Weathering Uncertainty" (Nakashima et al., 2012, pp. 29-31). [↑](#footnote-ref-144) 146. The spirit of the time, or “Zeitgeist” in German. [↑](#footnote-ref-145) 147. Many Indigenous peoples around the world have long understood and practiced the concepts of complexity and systems thinking through their traditional knowledge that sees how all parts of their local ecosystem connect and work together (Rasolt, 2021). For an excellent introduction to Indigenous thinking and its holistic approach to perceiving the world, we recommend Tyson Yunkaporta's book "Sand Talk" (2020). [↑](#footnote-ref-146) 148. See the R\&D agendas (UNDP Accelerator Labs, 2025b) developed in 2024. These agendas address key patterns identified across our network: the post-pandemic growth of digital finance and its unclear impact on poverty; the integration of grassroots knowledge with AI for sustainable food systems; and how businesses and governments can work together to create value from food and plastic waste. [↑](#footnote-ref-147) 149. For insightful discussions on bridging different knowledge systems, see the conversation with UNESCO's LINKS team on Indigenous and local knowledge systems (Knowledge and Learning GIZ, 2023) and "Making room for grassroots wisdom" (Rebello Fernandes & Mishra, 2024). [↑](#footnote-ref-148) 150. For a comprehensive overview of innovation approaches, see Nesta's Landscape of Innovation Approaches (Leurs, 2018b). [↑](#footnote-ref-149) 151. Leurs (2018a) [↑](#footnote-ref-150) 152. Based on Cohendet et al. (2010). [↑](#footnote-ref-151) 153. The concept of the middle ground was first introduced by Patrick Cohendet and colleagues (2010) in their research on creative cities. While they developed this concept to understand the situated creativity that emerges in such cities, we see similar middle grounds in sustainable development and other sectors as well. For us, the notion of a middleground serves as a lens that helps us see a space that is almost invisible at first glance, but once you see it, you immediately recognize its potential and value. It's important to note that in their original work, Cohendet and colleagues refer to the three layers as the upperground (institutions), middleground (R\&D space) and underground (grassroots). We have chosen different terms since their naming suggests a hierarchy that doesn't do justice to the ingenuity we see at the grassroots level. Also, the term “underground” can suggest subversive or illegal activities, while we actually aim to amplify and elevate what's happening at the grassroots level. [↑](#footnote-ref-152) 154. The role can be seen as a knowledge broker (Meyer, 2010), someone who facilitates the exchange of knowledge between different groups. They connect those who have knowledge with those who need it, translating complex information into accessible formats and ensuring knowledge flows effectively between different communities. But there is more than just brokering knowledge. The role of a boundary spanner is probably more accurate. A boundary spanner (see Tuschman, 1977) is someone who works across organizational and cultural boundaries to connect different groups and help them interact and collaborate effectively. They build bridges between organizations by understanding different perspectives, translating between different types of knowledge and ways of working, and building relationships that enable collaboration. [↑](#footnote-ref-153) 155. For a more conceptual perspective on shaping the middle ground, see Sarazin, Simon and Cohendet (2021, pp. 336-338). [↑](#footnote-ref-154) 156. Buchanan (2017) [↑](#footnote-ref-155) 157. Christiansen & Duggan (2021) [↑](#footnote-ref-156) 158. Kahneman (2011) [↑](#footnote-ref-157) 159. It may be worth exploring the "worldviews of change" that collectives hold, making explicit how each actor believes the world can and should change. Ingrid Burkett (2025) has developed a useful guide that could help facilitate such conversations. [↑](#footnote-ref-158) 160. See Alberti & Senese (2021, pp. 25-29), also see Crilly (2024) who discusses how different disciplines see, think and act. [↑](#footnote-ref-159) 161. In network theory actors are referred to as “nodes” (see Holley, 2013, p. 57). [↑](#footnote-ref-160) 162. For introductory reading on how networks work and how to build and curate them see "Impact Networks" by David Ehrlichman (2021) and June Holley's (2013) work on network weaving. [↑](#footnote-ref-161) 163. Holley (2013) [↑](#footnote-ref-162) 164. Altman & Nagle (2020) [↑](#footnote-ref-163) 165. Pfortmüller (2022); Mintzberg (2015) [↑](#footnote-ref-164) 166. See Romero (2018) for a very short summary. [↑](#footnote-ref-165) 167. See our SDG Commons platform: [↑](#footnote-ref-166) 168. Lucarelli (2023b) [↑](#footnote-ref-167) 169. Garud, Jain,& Tuertscher (2008) [↑](#footnote-ref-168) 170. Johan Redström (2008) refers to this form of design as “design through use”, instead of the traditional “use through design”. [↑](#footnote-ref-169) 171. Von Hippel (2001) [↑](#footnote-ref-170) 172. Chester & Samaras (2021). Take, for instance, the "half a good house" in Iquique, Chile, designed by Alejandro Aravena. In this project, families were provided with a basic structure, including essential plumbing and shelter, allowing them to expand and adapt the space as their families grew and their needs evolved (see Moore, 2016). [↑](#footnote-ref-171) 173. To create that fit Keith Grint (2010) suggests developing “clumsy solutions” which imperfectness allows to be more responsive to complexity. [↑](#footnote-ref-172) 174. Smith & Lewis (2011) [↑](#footnote-ref-173) 175. Smith, Lewis, & Tushman (2016) [↑](#footnote-ref-174) 176. In their 5-year reflection blog, UNDP Accelerator Lab Paraguay (2024) refers to this tension as "feasible rigor": generating robust evidence within real-world development constraints. Working with this tension actually enhances the process, as the Lab explains: "the tension between rigor and feasibility is unavoidable, but also enriching." [↑](#footnote-ref-175) 177. Mowles (2015) [↑](#footnote-ref-176) 178. Fernández Sirera et al. (2025) offer approaches for facilitating polarities in collaborative processes. [↑](#footnote-ref-177) 179. For practical guidance on navigating power dynamics in participatory work, see Hajira Qazi's "Power & Participation" (2018), which offers a reflection framework for change makers to evaluate participatory engagements and consider their own role and influence throughout the design process. Also recommended is Maya Goodwill's (2020) guide on power literacy, which helps develop understanding of how power and privilege affect collaborative work. [↑](#footnote-ref-178) 180. See our report "At the edge of participation" (Cottica, Sapienza, & Maarouf, 2025) for concrete examples of participatory approaches across diverse contexts. [↑](#footnote-ref-179) 181. Interaction Design Foundation (2021) [↑](#footnote-ref-180) 182. Tarabishy (2023) [↑](#footnote-ref-181) 183. For a brief introduction, see the “Portfolio Approach Primer” (UNDP Strategic Innovation Unit, 2022) or the report “Modernizing development: Introducing portfolios” (UNDP, 2025). [↑](#footnote-ref-182) 184. UNDP (2022) [↑](#footnote-ref-183) 185. Sternin & Choo (2000) [↑](#footnote-ref-184) 186. UNDP, GIZ Data Lab, & University of Manchester (2021). Also see our report “Learning from the Edges” (Pawelke, Glücker, Albanna, & Boy, 2022), where we document how the data powered positive deviance method helped us discover grassroots solutions using digital data. For an example of data powered positive deviance see Tchagam Mallam Boukar et al. (2022). [↑](#footnote-ref-185) 187. For more on data-powered methods that combines both traditional and non-traditional data, see Albanna et al. (2022). [↑](#footnote-ref-186) 188. While power and agency are related, they're distinct concepts. Agency refers to the capacity to act and make choices, while power determines whether those actions can actually influence outcomes. Someone might have agency (the ability to make choices) but lack power (the ability to make those choices matter). Conversely, someone might have power (influence over outcomes) but choose not to exercise their agency. See also: Agency. [↑](#footnote-ref-187) 189. We recommend Hajira Qazi's (2018) guide *Power & participation*, which offers a comprehensive framework for practitioners to reflect on their role and influence throughout participatory processes, with tools for understanding how power shapes who gets to participate and whose knowledge counts. [↑](#footnote-ref-188) 190. PowerCube (2011) [↑](#footnote-ref-189) 191. Birney (2021a) [↑](#footnote-ref-190) 192. Winhall, & Leadbeater (2021) [↑](#footnote-ref-191) 193. For further reading, see Maya Goodwill's (2020) *A Social Designer's Field Guide to Power Literacy*, which provides an accessible introduction to power dynamics alongside practical tools for mapping different types of power (access, role, goal, and rule power) to make design processes more inclusive and fair. Also see the Power Pack: Understanding Power for Social Change (Power Cube, 2011), which provides the Power Cube framework and practical tools for analyzing how power operates across different forms, spaces, and levels in social change efforts. [↑](#footnote-ref-192) 194. See Leurs & Duggan (2018) for a brief introduction into prototyping and how it’s different from other methods for testing solutions. For a more practical guide we highly recommend the prototype guide by Social Innovation Canada (see Cabaj, Tjennes & McNair, 2022). [↑](#footnote-ref-193) 195. See, for example, Lucarelli (2021) for the initial questions we developed to explore the dynamics between informality and digitalization. For the learning questions we developed on informal transportation, see the report by the Global Partnership for Informal Transportation (2022). [↑](#footnote-ref-194) 196. Verhulst, Chafetz & Fischer (2024) emphasize that questions are central to collective learning; asking better questions can help us create greater collective intelligence. Also see the work of Stefaan Verhulst (2023) on the “new science of questions”. [↑](#footnote-ref-195) 197. Kelly Duggan's (2019) blog “Think about it: Making the case (and space) for reflection” provides a concise and comprehensive introduction to the practice of reflection. [↑](#footnote-ref-196) 198. Schön (1984) [↑](#footnote-ref-197) 199. Mulgan (2018), Mulgan (2017, pp. 70-75) [↑](#footnote-ref-198) 200. Kolb (1984) [↑](#footnote-ref-199) 201. See Edmondson (1999; 2018) [↑](#footnote-ref-200) 202. Van der Bijl-Brouwer (2019a), also see Van der Bijl-Brouwer (2019b) for a more academic discussion. [↑](#footnote-ref-201) 203. For a brief but comprehensive discussion of the different purposes of framing, see Martinez (2021). [↑](#footnote-ref-202) 204. Pereverza (2022) [↑](#footnote-ref-203) 205. Busch (2020b) [↑](#footnote-ref-204) 206. Frameworks Institute (2020). [↑](#footnote-ref-205) 207. There are many more facets to relationships, of course. See the Relationship Project () for a more in-depth exploration. [↑](#footnote-ref-206) 208. Rye (2023) [↑](#footnote-ref-207) 209. For a comprehensive taxonomy of relationship categories within systems, see Birger Sevaldson's Library of Systemic Relations (2016). Also see his website: . [↑](#footnote-ref-208) 210. Winhall & Leadbeater (2022) [↑](#footnote-ref-209) 211. Orchard-Webb (2019) [↑](#footnote-ref-210) 212. Busch (2020, p. 242) [↑](#footnote-ref-211) 213. Based on Jason Pearman’s (2021) definition of Social R\&D. For more on Social R\&D see: socialrd.org and their report on “Social R\&D Practices and Patterns” (Social R\&D Community, 2019), or check out TACSI’s white paper (Curtis, et al., 2021). [↑](#footnote-ref-212) 214. See Gina Lucarelli's blog "The secret UNDP Accelerator Labs plan (just between you and me)" (Lucarelli, 2023b), where she describes our shift towards R\&D and outlines a plan for evolving the Network into an open, globally distributed R\&D capability for the Sustainable Development Goals. [↑](#footnote-ref-213) 215. Akinyemi & Leurs (2024) [↑](#footnote-ref-214) 216. For examples of collective R\&D in practice across diverse contexts, see our report "At the edge of participation" (Cottica, Sapienza, & Maarouf, 2025). [↑](#footnote-ref-215) 217. Mease (2022) argues that scaling needs to be more than just what's quantifiable, emphasizing that scale should be about depth, relationships, decentralization, and power rather than simply how fast something can grow or how many people a solution reaches. [↑](#footnote-ref-216) 218. This reflects the three main ways of scaling: scaling up (reaching more people and geographies), scaling out (influencing policies), and scaling deep (impacting cultural norms and beliefs) as described by Moore, Riddell & Vocisano (2015; also seeTulloch, 2016). For a practical example, see Mugema (2022) on how the UNDP Accelerator Lab in Uganda used this framework to inform their scaling strategy. [↑](#footnote-ref-217) 219. Tulloch (2018) calls this approach "scree-scaling," legitimizing and cultivating many small solutions rather than growing single ones, recognizing that system change is more likely to occur through the accumulation of many little ideas than a few big ones. [↑](#footnote-ref-218) 220. See Gupta (2016, p 341) [↑](#footnote-ref-219) 221. This approach of scaling through the ecosystem raises questions about who to scale with (see Mugema, 2022), and perhaps more fundamental questions, as Gord Tulloch (2018) suggests: Is scaling always the goal? If so, what should be scaled – and who decides? [↑](#footnote-ref-220) 222. For more: see the report “Transforming our world: The 2030 agenda for sustainable development” (United Nations (2015). [↑](#footnote-ref-221) 223. Busch (2024) [↑](#footnote-ref-222) 224. Busch (2020a, p. 14) [↑](#footnote-ref-223) 225. For more on weak signals and practical guidance on how to detect them and inform your work, see "How to Future" by Scott Smith and Madeline Ashby (2020), or the MaRS Startup Toolkit (Bolton, n.d.), or the Exploring Futures Guide by UNDP Accelerator Lab Argentine (Acosta et al., 2022). [↑](#footnote-ref-224) 226. Smith & Ashby (2020, p.69-70) [↑](#footnote-ref-225) 227. For examples of signal detection in practice, see: "The Changing Nature of Work” (Draskovic et al. (2021) where UNDP Accelerator Labs in Europe and Central Asia mapped signals of changing work patterns; the signal report by UNDP Accelerator Lab Argentina (2022b); Signals Spotlight (UNDP, 2024) and the collaborative signals platform at The Futures Centre (). [↑](#footnote-ref-226) 228. Complex skills (see Neelen & Kirschner, 2018) consist of many constituent abilities that interact with one another. And because they are interrelated, you can’t develop them by practicing one by one, they need to be developed in an integrated manner. [↑](#footnote-ref-227) 229. For a brief introduction, see Basma Saeed’s (2020) reflections on her position as the Head of Solutions Mapping in Sudan. For a practical guide, see the SalikLakbay Field Guide by the UNDP Accelerator Lab Philippines (Lor, 2021). [↑](#footnote-ref-228) 230. Von Hippel (2016); also see De Jong et al. (2023). [↑](#footnote-ref-229) 231. See Doorley (2012); Groves & Marlow (2016); Thoring (2019). [↑](#footnote-ref-230) 232. See the “Transforming our world: the 2030 Agenda for Sustainable Development” (United Nations, 2015). [↑](#footnote-ref-231) 233. See the report "Our Common Future" (World Commission on Environment and Development, 1987) – also known as the "Brundtland Report." For historical background on how sustainability was put on the global agenda, see Adam Rome's (2015) discussion of five key publications. For an in-depth analysis of the meaning, history, principles, pillars, and implications of sustainable development, see Justice Mensah's (2019) discussion of the concept. [↑](#footnote-ref-232) 234. See Donella Meadows' classic primer "Thinking in Systems" (2008). For more systems thinking resources and exploring the history and concepts of systems thinking, see The Systems Thinker journal's repository at thesystemsthinker.com. [↑](#footnote-ref-233) 235. We recommend reading Aarnout Wenneker's (2025) reflections on the limitations of systems thinking as a tool to navigate complexity and uncertainty. Traditional systems thinking assumes boundaries are clear, flows are predictable, and feedback loops can be managed with reasonable certainty about outcomes. But in complex systems, there is no control over how these systems behave or evolve: the whole isn't just greater than the sum of its parts – it can become something entirely different, shaped by interactions. As Wenneker puts it: "are you dealing with a system you can design, or one you can only navigate?" If it's the latter, we need complexity literacy alongside systems thinking. [↑](#footnote-ref-234) 236. See Sevaldson's (2016) classification of systemic relations. [↑](#footnote-ref-235) 237. This is what Leadbeater calls the "washing machine," (see Winhall & Leadbeater, 2020) a dialectical process where the old regime meshes with the new emerging alternative, similar to Cassie Robinson's (2019) version of the Berkana Two-Loop Model. [↑](#footnote-ref-236) 238. Adam Kahane (2025) suggests we should look for the cracks in a system and work with them. These cracks can appear as problems or as entry points to start transforming systems. They show systems aren't as solid as they seem, offering opportunities for fundamental change. [↑](#footnote-ref-237) 239. For more on tactics, see the Wheel of Trade Offs developed by Jesper Christensen (2019), a tool that may help you make decisions and align your lab strategy with your everyday tactics. [↑](#footnote-ref-238) 240. We recommend reading W. Brian Arthur's "The Nature of Technology" (2009) for his fascinating exploration of how technologies evolve through combination and recombination, continually opening new possibilities for innovation. Also worth consulting is Luke Jordan's guide "Don't Build It" (2021), which challenges practitioners to question whether building new technology is necessary before proceeding with development. [↑](#footnote-ref-239) 241. See Jeremy Boy's (2024) blog discussing the integration of data with technology and which configurations are best suited for certain development challenges. [↑](#footnote-ref-240) 242. Leurs & Roberts (2018, p 57) [↑](#footnote-ref-241) 243. See for example The DIY Toolkit (Nesta, 2014). [↑](#footnote-ref-242) 244. Heracleous & Jacobs (2011) [↑](#footnote-ref-243) 245. Tan (2023b). Also recommended is Vaughn Tan’s (2019) book “The uncertainty mindset”. [↑](#footnote-ref-244) 246. Christensen et al. (2017) [↑](#footnote-ref-245) 247. These ways of creating value through R\&D have been documented in the UNDP Accelerator Labs' R\&D Service Catalogue. See UNDP Accelerator Labs. (2025a, March). [↑](#footnote-ref-246) 248. The value chain represents a traditional concept of an industrial production process, where commodities flow in one direction through predefined stages (Normann, 2001). At each step, value is added until the process culminates in the monetization of products through market exchange (Vargo & Lusch, 2004). In contrast, value networks operate non-linearly, with value being created and exchanged through dynamic interactions between actors in an ongoing, continuous process (see Allee, 2003, pp. 192-208). Value networks can be understood as many-to-many systems (Dark Matter Labs, 2025): many actors connect in many ways to create and exchange many different forms of value. [↑](#footnote-ref-247) --- # Agent Instructions: Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter: ``` GET https://bas-leurss-personal-organization.gitbook.io/undp-accelerator-labs/references/glossary.md?ask= ``` The question should be specific, self-contained, and written in natural language. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.