Accelerating Innovation Through a Network of Ecosystems

What companies can learn from one of the world’s largest networks of accelerator labs.

Reading Time: 17 min 


Permissions and PDF Download
Accelerating Innovation

Image courtesy of Brian Stauffer/

A United Nations agency with a sweeping mission and sprawling global presence may not appear to be the most likely place where companies can learn new techniques for accelerating innovation — but appearances can be deceiving. The United Nations Development Programme (UNDP), an organization of 17,000 employees spanning 170 countries, focuses on solving the world’s most complex problems — ending poverty, ensuring healthy lives and well-being, providing affordable and clean energy, reducing inequalities, and more — through local, regional, and global initiatives and an annual budget of $5 billion.1 In our exploratory research,2 we have found that this organization is making exceptionally progressive innovation moves both at its headquarters and in the field by building and sustaining one of the world’s largest networks of accelerator labs — 60 of them, serving 78 countries.3

That’s why the agency provides such an instructive example for businesses. Usually, managers think about building an ecosystem for innovation, and most research-based advice focuses on challenges and opportunities that arise in that setting.4 But UNDP has gone much further, creating a large network of ecosystems that revolve around its labs, and building connections with local partners while also helping the labs coordinate with one another. In short, it’s solving the acceleration problem with a portfolio mindset, employing the power of the collective, and it’s doing so at an unprecedented scale.

Business leaders can take a similar approach to address innovation challenges across teams, units, and regions — and to tackle projects that would benefit from both local and global expertise. For inspiration, we’ll consider a few examples of multinational companies that have recently deployed their own networks of ecosystems. But first, let’s look at how the UNDP model evolved into what it is today and what it’s been able to accomplish so far. People tend to assume that all large government-related organizations are highly centralized, under-resourced, and sluggish, especially when it comes to trying anything new. However, over the past few years, under fresh leadership, UNDP has been learning from early mistakes and creating a new organizational structure to more effectively and efficiently ramp up its efforts in the area of sustainable development.

It’s a story of major change, to be sure. But similar outcomes are within reach for companies with bureaucratic obstacles to overcome as they try to build their own networks of ecosystems to accelerate innovation.

Evolution of a Portfolio Model

To derive management insight from the UNDP model, it’s useful to consider how it came to be and what initial innovation efforts looked like, some lessons learned from those, and the organization’s shift to a portfolio approach.

Early, fragmented local efforts. Despite the agency’s early tendencies toward risk aversion and bureaucracy, pockets of innovation existed in UNDP before it launched the Accelerator Labs initiative. Some were officially dubbed innovation labs, while others were small, independent teams acting on their own.

Generally, members of these groups weren’t hired as innovation people per se. Rather, they tackled innovation projects on the side. They often accomplished good things locally, but their efforts lacked broader impact because they were isolated, and so they weren’t successful in the eyes of UNDP leaders. For example:

  • In the country of Georgia, the government was using voice phone calls to alert citizens during disasters, which meant deaf people were not benefiting. So an innovation lab jointly operated by UNDP and the government brought the government’s disaster management team and people with disabilities together to cocreate solutions. Two of their ideas were prototyped and deployed: a text-based service and video calling to enable communication via sign language. Although the project addressed an important problem in practical ways, it went no farther than Georgia.
  • In Egypt, a lab organized a five-day social innovation camp to rally young people to improve the economic situation of fishermen in sustainable ways. Young women and men worked with local innovation groups and the Ministry of Local Development to develop solutions such as technologies to improve fish stocking, new designs for recycling fish waste, and efficient tools for shrimp peeling. But because those solutions stayed local, the lab’s work wasn’t perceived by UNDP leaders to have the kind of impact that would be appreciated by ministers of employment or finance.
  • In Croatia, employees in the local UNDP office sought money for solar panels to generate energy for children’s schools. The team worked with local leaders to adopt a new crowdfunding model. It worked, but the effort to promote crowdfunding for projects in other regions stalled because it fell outside mainstream UNDP funding processes.
  • For tuberculosis patients in Moldova, a team launched a two-year randomized trial to improve drug administration adherence. It worked with a telecommunications provider so that doctors could use Skype to see patients taking medications. The project achieved its goals, and the team even received external funding. But UNDP didn’t broadly adopt the methodologies because the team was neither aligned with head office priorities nor visible to management.

In these regions, entrepreneurial UNDP employees embarked on reasonable projects by identifying clear problems, developing innovative solutions (often with local partners), and implementing them. Yet they didn’t create enduring change elsewhere, nor did they address the large-scale multidimensional problems that are core to UNDP’s mission.

Lessons from those initial projects. As UNDP leaders designed their new Accelerator Labs network, they incorporated the following four lessons gleaned from those early efforts:

  • Align project choice with broader goals and activities. In some cases, the early teams embarked on experiments unrelated to anything else underway in their local offices or across the organization. They invested in innovation projects, developed new skills, collaborated with government agencies, and even found new funding. But since they didn’t build on existing organizational priorities, they didn’t change the way UNDP worked. Their projects were seen as one-off or ad hoc.
  • Stay visible to management. Initially, the innovation labs and teams believed it would be beneficial to stay under the radar, both in their local offices and in the larger UNDP organization, so they could test new approaches without encountering political roadblocks. They followed long-standing advice that suggested innovators should separate themselves from the mainstream as they experimented.5 However, this strategy backfired because they weren’t close to power centers in their organizations. They would have done much better to remain visible to local management or headquarters (ideally both).
  • Connect with internal experts. The innovation labs and teams felt that their mandate was to bring new ideas and knowledge into the organization. Thus, they consulted with external specialists but often neglected experts inside the organization. Even though U.N. offices had relevant expertise in disaster risk, governance, development, and other areas, innovators ignored it, fearing that internal feedback would stall them. However, not surprisingly, weak integration between new and existing U.N. operations hindered adoption of the new efforts. Experts in U.N. offices resisted new solutions because they hadn’t been given an opportunity to vet them, contribute ideas, or participate in their implementation.
  • Coordinate across geographies. The original innovation labs and teams were geographically isolated. They weren’t part of a larger network that could support sharing, cocreation, and learning. This resulted in weak or nonexistent integration across cities, countries, and regions. Essentially, even though UNDP was a global organization, leaders realized that they weren’t reaping the benefits of scale and scope.

The lessons applied: Shifting to a network of ecosystems. In 2018 and into 2019, new leaders at UNDP recognized that they needed a better approach to innovation and organizational learning to accelerate their impact on sustainable development. So they decided to build a network of accelerators across the globe. The Accelerator Labs interact with businesses, local governments, citizens, schools and universities, other NGOs, and their local UNDP offices to accomplish their mission of speeding up learning and execution while also searching for, creating, and sharing new solutions. Further, the labs tap into their collective intelligence to support their own learning.

Drawing on their understanding of how early isolated efforts fell short (the four lessons described above), the Accelerator Labs now (1) investigate what is already in a local office portfolio, (2) work with management to determine where they can contribute, (3) leverage UNDP internal experts residing outside the Accelerator Labs, and (4) collaborate with partner organizations within and across geographies — all while exchanging insights with the other Accelerator Labs around the world.

At the Zimbabwe lab, for instance, we’re now seeing projects that sync up with other UNDP activities and goals. The UNDP office in Zimbabwe is promoting resilience in rural areas by working with the World Food Programme to improve conditions for people who barely have the resources necessary to live. The country’s Accelerator Lab has chosen to expand the focus of this work to promote resilience and reduce risk also in urban areas that present equally severe but distinct challenges. This effort aligns with the priorities of the UNDP director and the Zimbabwean government, while enabling the lab to create and use a new ecosystem to solve a complementary problem. For example, in November 2019, the UNDP worked with the Urban Resilience Programme, UNICEF, and the Ministry of Local Government Public Works and National Housing to organize a hackathon to develop solutions to address economic and environmental challenges experienced by urban populations in Zimbabwe. What’s more, to take advantage of (and increase the effectiveness of) the Accelerator Labs network, the Zimbabwe lab recently coordinated a meeting, called Harare Innovation Days, with representatives of 35 other UNDP offices on the African continent to discuss issues and share best practices.

The lab in Serbia provides a good example of staying visible to management. Because depopulation is an enormous problem in the country — the population has decreased 5% in the past decade and is expected to drop another 19% by 20506 — the Serbian government has established an emergency task force to address it, and the new UNDP regional director has committed to work with the group.7 This is a multifaceted problem and a top priority for the nation — the type of problem a UNDP innovation lab previously would have avoided because of the complexity and visibility involved. But the new Accelerator Lab is actively engaging with it, using an ecosystem approach and working with local stakeholders. To start, it is analyzing a data set recently published by the World Bank Group from LinkedIn’s professional network. The goal is to study labor dynamics across countries, industries, and skills, get a clearer perspective on the problem, and share insights with local government ministries.8 Grappling with a similar problem and leveraging the Accelerator Labs network, the UNDP team in Bosnia has begun working with the Serbian team to learn from its efforts.

In Vietnam, the Accelerator Lab team is embracing the concept of becoming more integrated with the existing local UNDP organization, working with internal experts, and aligning with office priorities — for instance, its interest in new projects related to medical plastic waste.9 The Accelerator Lab is also cooperating with the local UNDP office by mapping the external ecosystem of actors involved with these efforts, including groups and entities UNDP would not traditionally engage with, such as informal waste pickers, local startups, a local engineering school, and people with related interests, such as bicyclists who care about having a cleaner environment. The lab team will also be reaching out to those who bring new financial tools and think differently about how to tackle such a complex, multifaceted issue like waste management.

Exemplifying the effort to become more geographically integrated and communicate across the Accelerator Lab network, the Lebanese Accelerator Lab is sharing best practices with the Accelerator Lab in Colombia. Lebanon changed the way it does employee recruiting and is using more modern group and case interviewing techniques; Colombia heard about the Lebanese hiring process and sought to adopt similar techniques.10 While Colombia didn’t have the time or resources to apply the same process, it was able to use videoconferencing technology to modify the system to work within its local constraints.

Across the UNDP, the Accelerator Labs are continuing to build ecosystems to solve local problems while also actively engaging with one another across their worldwide network. This effort differs from what we sometimes see with distributed R&D units in three primary ways. First, the labs are sparsely populated, with only three staff members each (see “Fewer Experts in More Locations”), so they are forced to engage externally. Second, in part because of their scant resources, the labs work with and learn from one another to expand their influence. And third, the Accelerator Labs often seek to accelerate innovation efforts already underway rather than find entirely novel problems and solutions. While this is a subtle shift from traditional innovation mandates, it is profound in that it aligns the labs’ efforts with prevailing organizational goals.

Takeaways From Company Experiences

In our research, we found a number of companies pursuing a network of ecosystems model. The appeal is clear: As the costs of internal innovation continue to rise, businesses are increasingly looking for external sources of innovation.11 Some recognize that a network of ecosystems can help harness distributed innovation, but those that have taken steps toward adoption are for the most part still in their early days. They can learn a great deal not just from UNDP’s efforts but also from companies that are trying similar techniques.

Some businesses have established full offices in innovation centers like Silicon Valley or near research university hubs like the Research Triangle in North Carolina. That approach can certainly yield benefits such as local knowledge spillover and ease of hiring qualified staff. But the leaner UNDP model is much more efficient. A company that convenes, leads, and joins ecosystems in multiple regions becomes highly embedded in and engaged with local problem-solving while limiting the need for larger-scale investments. When satellite innovation offices become ecosystem orchestrators in their regions, they act as connectors, instigating and accelerating projects. Meanwhile, they keep headquarters apprised of what is happening locally so that the main office can facilitate valuable connections with other satellites while also enabling the satellites to communicate with, learn from, and collaborate with one another directly, making the most of the network.

Metaphorically, like the UNDP Accelerator Labs, these satellite offices are the suns of their own ecosystems, but they’re also part of a much larger universe, not simply extensions of the company’s R&D organization. That universe enables them to solve complex, externally driven problems, use local lessons to apply or tailor innovations elsewhere, and diffuse novel solutions.

Adopt a network approach to solve complex, externally driven problems. Because many problems that countries face have sources outside their own regions, UNDP has found that it makes sense to solve them with a network approach, often drawing upon lessons from outside regional borders. Portfolios of efforts in various countries connected by similar themes and external drivers make the overall organization smarter.

An example of a company applying this same logic to solve complex, externally driven problems is Syngenta, a Swiss agribusiness company. Syngenta has developed a network of digital innovation labs in the U.S., the U.K., India, and Singapore to tap into local efforts to digitally transform agriculture. Similar to the UNDP Accelerator Labs, the newest Syngenta lab in Singapore is sparsely staffed; it employs only six people. This limited head count encourages the lab to partner with local innovators, fostering a culture of experimentation while still allowing the company to scale its innovations.12 One of the digital innovation labs developed a mobile app that lets farmers take a picture of an unknown insect or disease affecting their crops, then identifies it and provides information on potential solutions, including Syngenta products. This app would not have been developed if Syngenta did not have employees in local areas interacting with customers, and it has since been scaled across the organization to deal with crop infestations across the globe.

Think expansively, and use local lessons to apply or tailor innovations elsewhere. Perhaps one of the most difficult aspects of executing a network-of-ecosystems approach is learning from local solutions, generalizing them, and then (where appropriate) relocalizing them for other geographies. This process of innovating from the edges can be complicated because even if two locations face the same problem, local geography, culture, regulations, and other norms may limit the ability to implement the same solution in both places. Therefore, to generalize and relocalize lessons learned, the satellites must be in close touch with both the local stakeholders (to understand their demands and needs) and the parent organization (to understand the limitations of existing solutions and technologies).

A good business example is the global network of innovation centers developed by pharmaceutical and agriculture company Cargill. Although Cargill’s centers are larger than the UNDP Accelerator Labs (Cargill’s have a very broad mandate, and they each house customer support centers), the centers are still closely linked with one another, which enables them to share regional trends and requirements they’ve identified locally. That shared knowledge allows Cargill to develop versions of innovative solutions that will work in multiple regions or even one version that will work across regions.13 For example, when Cargill developed a new low-calorie sweetener based on the stevia leaf, the company faced a wide variety of regulations in different countries. Instead of developing completely different sweeteners for each market, it tapped into its local innovation centers to understand the limitations imposed by each country’s regulations and then set out to conduct research and product testing to satisfy all of the regulations with a single product. In this way, Cargill avoided the costs of inventing different products for each market.14

Diffuse innovations throughout the network and beyond. Even when valuable innovations occur, they often don’t spread until long after they are created. It’s a long-standing problem.15 Whether the innovations are products or processes — and whether they originate inside or outside the company — a network of ecosystems can be used to accelerate diffusion from one location to another. For example, AT&T’s Foundry connects the company with more than 500 startups around the world via local offices to identify innovations that could benefit AT&T’s customers, augment AT&T’s own innovations, and help bring new technologies to the market faster.16 By streamlining the engagement process and onboarding partner startups into the ecosystem, AT&T enables entrepreneurs across the globe to have quick access to the company’s resources and expertise to codevelop technology and prepare it for delivery. For example, AT&T Foundry has a team in Israel that is constantly on the search for local innovations that it can help develop and then diffuse to its partners worldwide. This is how it came across Vorpal, a 12-person company developing a method to detect, locate, and track drones to help protect airports and other critical assets. Vorpal teamed with AT&T, allowing it to access AT&T’s edge computing environment so the company could grow more rapidly, and then AT&T helped diffuse the technology to airports across the globe.17

Although business leaders may not think their company has much in common with the U.N., many of UNDP’s former limitations (a large, slow-moving, global organization with a great deal of formal authority in its blood) will sound familiar to managers in companies where the processes for identifying, accelerating, and diffusing innovations are sluggish. We believe such organizations can take a page out of the UNDP playbook and build a network of ecosystems that are individually effective yet collectively even more powerful to address large-scale challenges using insights gained from localized innovation and expertise.

Despite the many benefits of using a network of ecosystems to harness dispersed innovation, there are numerous challenges. In particular, it can be difficult for organizations to hire appropriately for ecosystem development and management, shift organizational leadership and culture to a much more open approach to collaboration, and strike the right balance between top-down direction versus bottom-up adaptation. However, if local ecosystem teams, working in support of larger organizational goals, are able to align project choices, stay visible to management, connect with internal experts, and coordinate across geographies, they can harness the power of this model to solve complex problems, apply or tailor innovations, and diffuse solutions both across their organizations and more broadly.

Editor’s Note: This article appears in the Summer 2020 print edition under the headline “Accelerating Innovation.”



1. “Sustainable Development Goals,” UNDP, accessed Dec. 21, 2019,, and “UNDP Strategic Plan, 2018-2021,” United Nations, Oct. 17, 2017,

2. We have met with leaders of UNDP, participated in a UNDP networked-learning workshop, and reviewed UNDP and third-party reports, articles, and blog posts to understand the agency’s recent efforts.

3. “Accelerator Labs,” UNDP,, and “UNDP Accelerator Labs,” Medium, accessed Dec. 21, 2019,

4. See, for example, this classic article: R. Adner and R. Kapoor, “Value Creation in Innovation Ecosystems: How the Structure of Technological Interdependence Affects Firm Performance in New Technology Generations,” Strategic Management Journal 31, no. 3 (March 2010): 306-333. For a more recent review of the subject, see R. Kapoor, “Ecosystems: Broadening the Locus of Value Creation,” Journal of Organization Design 7, no. 1 (2018): 1-16.

5. C.M. Christensen, “The Innovator’s Dilemma: When New Technologies Cause Great Firms to Fail” (Boston: Harvard Business School Press, 1997).

6. G. Filipovic, “Serbia Tackles the Plunging Population Plaguing the Balkans,” Bloomberg, Feb. 26, 2019,; and A. Kiersz, “The 20 Countries Facing Population Collapse,” Business Insider, Jan. 17, 2020,

7. F. Pickup, “How to Address a Shrinking Population,” UNDP, Oct. 22, 2019,; and I. Cerovic, “Depopulation: What’s It All About?” UNDP Serbia, Dec. 31, 2019,

8. D. Draskovic, “A Glimpse Into LinkedIn Data to Understand Serbian Labour Out-Migration,” UNDP Serbia, Oct. 8, 2019,

9. “Mitigation of Medical Plastic Waste in Health Sector in Viet Nam,” UNDP Viet Nam, Aug. 16, 2019,

10. For a sample Accelerator Labs job posting, see, accessed Jan. 2, 2020.

11. See, for example, C. Baldwin and E. von Hippel, “Modeling a Paradigm Shift: From Producer Innovation to User and Open Collaborative Innovation,” Organization Science 22, no. 6 (November-December 2011): 1399-1417; M. Bogers, H. Chesbrough, and C. Moedas, “Open Innovation: Research, Practices, and Policies,” California Management Review 60, no. 2 (winter 2018): 5-16; and E.J. Altman, F. Nagle, and M.L. Tushman, “Innovating Without Information Constraints: Organizations, Communities, and Innovation When Information Costs Approach Zero,” in “Oxford Handbook of Creativity, Innovation, and Entrepreneurship,” eds. C. Shalley, M. Hitt, and J. Zhou (Oxford: Oxford University Press, 2015).

12. A. Soh, “Syngenta Opens Digital Innovation Lab in Singapore,” The Business Times, April 18, 2018,

13. “Research & Development Facilities,” Cargill, accessed Dec. 27, 2019,

14. “Substantiating the Safety of a New Sweetener: Regulatory & Scientific Affairs Case Study,” Cargill, accessed Jan. 17, 2019,

15. For example, it took nearly 50 years after the discovery that citrus fruits helped prevent scurvy for the solution to start diffusing widely. See M. White, “James Lind: The Man Who Helped to Cure Scurvy With Lemons,” BBC News, Oct. 4, 2016,

16. “Who We Are,” AT&T Foundry, accessed Dec. 27, 2019,

17. “Startup Nation: Accelerating Innovation,” AT&T Foundry, accessed Feb. 7, 2020,

i. See post explaining three new roles: “UNDP Is Hiring Creative Talent to Join a Groundbreaking Network of Accelerator Labs in 60 Countries to Re-imagine Development,” UNDP Azerbaijan, March 4, 2019,


The authors would like to thank the United Nations Development Programme and the leaders of its Accelerator Lab initiative for providing background information for this article and for working to solve some of the world’s most complex and important problems.

Reprint #:


More Like This

Add a comment

You must to post a comment.

First time here? Sign up for a free account: Comment on articles and get access to many more articles.