Accelerator Labs: the challenge of engaging the mothership

UNDP Accelerator Labs
11 min readMar 14, 2019


A tale of 3 moves

At a time when many would claim that we reached “peak labs”, UNDP is launching 60 Accelerator Labs across the world. How do we build on lessons from other organisations and our own innovation journey to prevent the new labs from becoming (yet another) island, separated from the core business at UNDP? In this post, we share our current thinking, based on the hypotheses that the best way for labs to accelerate impact is not to create new projects at the edge, but to engage with what is already in place (the existing project portfolio) and provide a new frame to understand patterns and relevance gaps. This takes us into the territory of portfolio sensemaking and portfolio design, and creating new mindsets and models to think with, rather than methods and innovation pilots.

By Giulio Quaggiotto, Alexandru Oprunenco, and Bas Leurs

Chris Barbalis on Unsplash

It is an exciting — and daunting — time to be working on innovation at UNDP. After years of coy embraces, innovation is now at the forefront of the organisation’s ambition to find a new relevance. Early innovation champions sporting the scars from years of guerilla warfare find themselves in the unusual and somewhat uncomfortable situation of being all of the sudden “mainstream” and under the spotlight. At a time when high profile public sector labs are closing, UNDP has taken the bold step of launching 60 accelerator labs across the globe. The stated intent is to build “the fastest learning network around development challenges”. No pressure!

In the Asia Pacific region, we are interpreting the renewed mandate for innovation as an opportunity to reframe: building on insights from past efforts whilst at the same time gathering energy around a new intent. As Keren Perla memorably put it “start-ups like Labs must always have at least one eye on what their next move might be”. In our case, our sight is now turning towards the “belly of the beast”. To find a new relevance and accelerate impact, this is the hypothesis we are exploring, we need to develop a new engagement mechanism with the mothership: the way UNDP understand itself and builds its rationale for action. Whilst in our early innovation efforts the emphasis was on labs as places at the edge that can test and prototype new approaches that can be scaled (both internally and externally), in this new phase we are pursuing a more ambitious, perhaps impossible, but arguably more long-lasting line of exploration: changing the rationale under which the organisation organises its project portfolio and learns from it.

Three innovation trajectories: (a) at the edge of the organisation, (b) engaging with the mothership, and (c) building portfolios of learning options

For this reason, after having resisted for a long time attempts at defining innovation, for this new phase we decided to adopt Luca Gatti’s words: “innovation is furnishing a system with renewal capability”. This focuses our efforts on institutional innovation. As Esko Kilpi reminded us, an organisation emerges out of patterns of relations and decisions making. Hundreds of decisions to do (or not to do) things differently are taken at all levels across an organisation (and its partners) every day. They determine its ability to innovate and learn. It is those patterns that we are interested in surfacing and eventually shifting to answer more effectively questions of relevance and coherence. This lies at the heart of our promise of becoming “the fastest learning network” in development. It is the renewal capability that we ultimately want to bring to governments and partners to accelerate impact and open new possibilities at a time when collective imagination seems to be locked into the search of unicorns and quick fixes.

First move: of antibodies and centrifugal forces

To oversimplify, then, our current pivot can be described as the transition from a largely centrifugal to a centripetal trajectory. It is a trajectory similar to the one described in a magisterial post by Dan Hill when he talks about moving from user centered design to strategic design:

“In Argentina recently, a digital government-led service design approach envisioned, designed and delivered digital driving licenses in 65 days, from scratch. This is extraordinary compared to the comparatively geological pace typically witnessed in policymaking and delivery. But could such processes radically reduce the number of people driving in the first place?” Dan Hill

For us, going “centripetal” means asking more fundamental questions of the “should we be driving in the first place?” type (without, of course, neglecting the importance of doing that got the innovation agenda running in the first place).

But our first generation innovation efforts were — by necessity in a context that was not necessarily open to exploring new approaches — very much focused on the grassroots: surfacing early adopters and creating a political cover for them so that the organisation would not squash their enthusiasm. Setting up labs, training cohorts of innovation champions and making seed funding available for experimentation in this context were tactics to create a state of exception on the inside and generate interest from the outside. The results in many cases exceeded expectations, as the recent launch of Citra Lab by Sri Lanka’s Prime Minister reminded us. But overall it is fair to say that the trajectory of these efforts has been centrifugal. The mothership’s antibodies reacted against attempts at injecting a new way of thinking and doing, and what started at the edge has been largely pushed even further out, often encountering more validation outside than within the organisation. Lots of pilots have been initiated, but they did not result in decisions to do things differently nor the ambition to bring about system level change (a recurrent challenge, as Jesper Christiansen et al. reminded us recently). As a result, the impact has been limited. The organisation’s innovation muscle has undoubtedly been built in some pockets and a subterranean vein of bottom up energy has been unveiled. These are important results. But the ambition of changing mindsets and habits across the board has not quite materialised.

Second move (inwards) — focusing on the core

The new phase of our innovation journey has brought a new frame to it — namely, relevance. What we realised is that for many of our country or regional offices, it is still quite difficult to answer two fundamental questions:

  1. Why do we have this particular set of projects at this point in time?
  2. Are our solutions relevant and coherent with the complex problems faced by the country in which we operate? (This is not too far away from a government having to make sense of its current set of policies, say, on climate or future of work as they are dispersed across different ministries and often focused on “quick fixes” rather than systemic solutions).

Without being able to provide a new perspective on these questions, the innovation agenda might be confined once again to the margins. Hence the focus on a centripetal trajectory. Our current hypotheses is that one of the biggest services that the newly established accelerator labs can provide to the mothership is not to start developing a whole new set of experiments at the edges (at least to begin with) but rather to take a different look at what is already in place at the core (the project portfolio) and delve deep into questions of relevance and coherence. It is the equivalent of taking a look from the balcony to what is happening on the dance floor, and in the process allowing for a new process of seeing to emerge.

This means taking the “acceleration” metaphor quite literally: accelerating the impact of what is already in place, rather than waiting for the generative, R&D process of labs to take its course. Importantly, the underlying mechanics of generating acceleration momentum is one of layering and aggregation, rather than elimination by funnelling (trying many things in the hope that one of them will stick and scale to unicorn level).

For the reason illustrated above, at this particular point in our innovation trajectory, this translates into a focus on portfolio sensemaking and design (we are somewhat comforted by the fact that we are not alone in exploring portfolio approaches to development challenges — some recent examples include GAHI, PATH and Climate Kic)

Second move (upwards): abstraction and creating new meaning

“This is often what we mean by creativity: seeing in new ways, spotting patterns, and generating frames. ” Geoff Mulgan

Our attention was drawn to a particular approach to sensemaking, developed by Axilo. We are now working with them to adapt the protocol to a development context, embed it in our accelerator labs, and, eventually, offer it to our government counterparts. In practice, this would mean that the first step in our engagement process with UNDP programs and partners will be to offer the capability of looking at an existing set of projects and activities, separate the noise from relevant patterns and signals, create new frames and extrapolate actionable intelligence from them. Note that in this process the word “innovation” is hardly ever mentioned — yet another notable difference from our engagement model in the past.

Early draft architecture of UNDP sensemaking protocol

There are a number of reasons why we were drawn to this approach to sensemaking in our particular context:

  • Many of the foresight or system mapping exercises we have been involved with in the past often fall short of practical insights into what to do next, once the complexity and interrelated nature of the challenges we have to face has been recognised. The emphasis on sensemaking, as an eminently social and iterative process, partially compensates for these shortcomings. This is not about a consultant or a small group of experts producing an analysis of the current portfolio and then presenting recommendation for action, but, in our case, a whole office (and external partners) having a shared, recurrent experience of reflection and extraction of meaning from a current set of activities by focusing on patterns. If the process is managed successfully, a group can move from “first loop learning” (learning from within an existing paradigm) to creating new categories and models to think with (“second loop learning”, as per Geoff Mulgan’s definition)
Loops of intelligent learning (Source: Geoff Mulgan)
  • The clear distinction between sensemaking as a process and actionable intelligence as an outcome. The latter zeroes in on those decision patterns that we are interested in shifting. If sensemaking allows options and possibilities for renewal to emerge, someone, somewhere, at any level within the organisation has to take a decision to accept, reject or reframe those options. An argument for change has been made, based on learnings from experience. Someone has to respond to that argument. The decision is under the spotlight. This is where learning effects can be accelerated: no more pilots, but genuine experiments, in a portfolio logic. This is also where we can build capacity and organisational muscle to produce compelling arguments for change.
  • The relentless focus on the “why” (“can we get less people to drive?”) rather than the “what”, or “how” of existing projects (“can we make it easier for people to get a driving licences) — and how this relates to the stated intent of the organisation. Being able to reflect on the difference (if any) between original intent and how this was translated into projects on the ground is ultimately a mindset shift
  • The networking effect generated as collective reflection “connects the dots” across organisational silos, allowing for stronger coherence to emerge. For instance, in one of the first field sense-making experiences we discovered that operation of e-commerce centers in rural Bangladesh has not only led to more dynamic economic life there (an anticipated 1st round effect), but also through women empowerment as they have started e-shopping (unanticipated 2nd round effect). This opens the way for a more systemic and concerted programming but also help teams rethink their portfolio from systems’ change perspective.
  • The shift from portfolio as a management/accountability tool to a learning tool. Ever since Panthea Lee’s seminal post on “downstream vs upstream” data, we’ve been keen to distinguish reporting from learning. With sensemaking we are de facto prototyping a system focused on accelerating learning (whilst most of our current systems are geared towards reporting, for obvious, and important, reasons).

It is way too early for us to draw any insight as to whether our hypothesis that sensemaking is the most effective mechanism to engage with the mothership is valid or not. Our early tests on the ground in Bangladesh (with the A2i team) and Cambodia (with government counterparts) have highlighted that building abstraction, pattern recognition and intelligence extraction skills across a broad constituency is no minor challenge (more in an upcoming post). At the same time, these skills seem to be foundational if we want heavily projectised organisations not to lose track of the bigger picture and move up the learning spiral. We will no doubt learn more as we are rolling out the protocol to countries across the region — and beyond — in the upcoming months.

Sensemaking session in action in Bangladesh: identifying patterns to create new frames

Third move: building portfolios of learning options

“The important thing for government is not to do things which individuals are doing already and to do them a little bit better or a little bit worse, but to do those things which at present are not done at all” R. M. Keynes

Of course, sensemaking and intelligence are not an end in themselves. And there’s only that much acceleration impact that can be derived by layering and aggregating what is already there. If we want to live true to our promise of being the fastest learning network on development challenges, we will need to develop the capability of building portfolios of learning options around topics that are new both to us and the governments we are working with. Take for example a question like the future of work. Analysis and predictions from experts and consultancies in this area are endless, and endlessly diverging. Governments are sold a variety of quick fixes: from teaching everyone to code to blockchain for traceability. The paradox of the obsession with single point solutions is that rather than providing more options in face of complexity, they reduce them.

Why not, instead, work with interested governments to build a portfolio of interventions designed to accelerate learning and provide future options, before committing to a particular policy intervention? The ability to design the portfolio so that if provides sufficient diversity and velocity, combined with the ability to manage multiple parallel experiments will become, we believe, key differentiating factors for those organisations that are interested in creating new possibilities and bring about systemic change in the face of wicked development challenges.

The third “move” then takes us back to the centrifugal trajectory — away from the internal logic of portfolio composition and outwards to explore new areas, investing in learning options together with those organisations and communities that know much more about a particular challenge than we do.

We will keep sharing our learning as our journey toward a new relevance continue. In the meantime, all comments and suggestions are welcome! Do get in touch.

With endless thanks to Milica Begovic and Dias Rahwdiati for their comments



UNDP Accelerator Labs

Building the world’s largest learning network around development challenges. 91 Labs in 115 countries.