Locations
UNDP is making a large bet on the lab model in order to create space for experimentation and drive learning on what works in the face of rapidly changing development problems. This is an initiative about making space for creativity in the face of problems that need new methods & new energy. For us it also means disabusing ourselves of the notion that there are one-shot panaceas for the world’s development problems.
As an exercise in liberation, setting a predefined set of indicators of success for UNDP’s Accelerator Lab initiative has been tricky. In my experience, innovators tend to avoid setting targets because the nature of innovation work is meant to break new ground and create new models. Ex-ante target setting won’t help you count the things that haven’t been invented yet. (It’s the classic “who-knew-anyone-needed-an-I phone-till-we-all-had-one” argument…)
The very mention of results-based management raises cackles among innovators in the development space. Maybe because many innovators are cognizant of the limits of centrally driven, linear planning within complex adaptive systems. Or maybe it is because, well, who likes to be put in a (logframe) box? Still, we need to know what success looks like.
Our version of directed improvisation: Yes, no,….we don’t know.
The team of designers behind the Accelerator Labs (@ElaMi5, @bbusetto, @jd_dcruz and the late @gmehn) wanted to create conditions for adaptive, bottom up improvisation. Building on Yuen Yuen Ang’s maxim “directed improvisation” from her book How China Escaped the Poverty Trap, the core idea is that the center directs and the front line improvises. Ang’s work on directed improvisation has inspired us to set boundaries of what must be done (green space), clarify boundaries of what cannot be done (red lines) and leave a large gray space in between.
When we send off our Accelerator Lab teams to make change out there in this messy world, we try to make direct improvisation literal. We draw this red line circle on the floor and a green spot in the centre. We tell the lab teams something like this: have a look at the green spot, this is what you must do. The only thing you must do is try to solve development problems and accelerate learning on how to solve complex challenges. The do-not-cross red line — drawn in literal red tape — signifies the things you must not do like commit financial fraud, steal community intellectual property, hurt people during experimentation and the like. The rest? It’s up to you. That leaves a huge gray area where instead of issuing edicts, we will follow the lab network’s way of working as it emerges in practice.
Public sector lab success: by culture, process and/or product?
Outside of the development space, there’s lots to learn from many public sector innovation labs. What success looks like focuses on the process and culture changes that are necessary for innovation and renewal. The States of Change’s cultural change impact framework is instructive. The framework is useful as it shows the gamut of changes needed to address complex challenges (attitudes, abilities, behaviors, roles relationships etc.) One can imagine using this frame as a way to see the ripple effects and indirect value creation as a success measure in labs.
Spinning off from that, John Kenny, who works in a government innovation lab at the federal level in Canada, presents a similar framework to demonstrate the value of their innovation team. This framework converges with the States of Change impact framework in some ways.
Both tools recognize that labs should be assessed both by the outputs they create (new products and/or ways of working) as well as the softer side — attitudes, mindsets, relationships, and even how things are discussed. From this work the take away seems to be: for innovation to take hold, the intangible bits are part of what lab success looks like. This might be a hard lesson for those of us programmed for results, results, results.
The (ongoing) reinvention of the Logframe
Speaking of results …our worldwide lab experiment is unfolding in a development context that operates within results-based management. There’s no alternative (yet). Still, there are signals of a broader questioning of the assumption that stewards of public funds can forecast exactly what effect an intervention will have. Logical frameworks (the result-based management tool of choice) depend on linear paths, and anyone who’s ever tried to do something ambitious in a volatile, multi-stakeholder environment has doubts that they can ensure we do “a”, it leads to “b” and then “c” goodness in the world happens.
In other words, as Mike Tyson said: “everyone has a plan until you get punched in the face”. For those working towards sustainable development, we plan an awful lot, get punched in the face and then we tend to justify in reports all other unplanned good things that happened along the way that we couldn’t anticipate at logframe stage.
A change is coming…maybe. There are hopeful signs that iterative, agile-style management is infiltrating development. The search frame as a successor to the linear logframe that Duncan Green refers to in his blog is one such example that can be instructive for our work with Accelerator Labs. Here, it’s not a uni-directional line that is drawn — “if we do this, then this observable sign of progress can be attributed to our efforts”. Rather there are built in iteration checks which allow for multiple paths to potential success.
This way of thinking aligns with the “Futures Wheel” we have been using within the UNDP Accelerator Lab bootcamps, helping the teams to explore the future implications of their challenge. What if we stretch this idea to logframes?
Maybe instead of asking what exactly is the change we expect to see from our investments and experiments, perhaps the better question is: what would a logframe look like if it had multiple possible outcomes — some which are considered the desired effect and some which are unintended consequences? What if logframes were radial rather than linear expressions of the theory of change?
Traditional logframes plot out a linear pathway for a desired future, versus a radial logframe, considering multiple possible outcomes: desired, unintended (but positive) and unfavourable.
Early thoughts: Three signs of a successful lab
Hopefully I’ve hedged enough against traditional KPIs for the innovators out there to trust me. Either way…rolling up one’s sleeves, and in the chillest possible way…what are the signs of success for a lab meant to accelerate learning and problem solving in sustainable development? Here’s some early unframed thinking on three directions for knowing a successful UNDP Accelerator lab when we see one:
Diversity to tackle complexity
Much of what we are trying to do is to diversify how we understand problems and solutions. Here we are inspired by the Playbook for Collective Intelligence and Dave Snowden’s Seven Principles. Taking on complex development problems requires acknowledging at least two things about reality: it changes all the time and it is distributed among many actions and choices. This means if you acknowledge that sustainable development is a by the minute type of problem, to tackle it we need to diversify where we get our intelligence.
What would this look like? It could be that we help decision makers learn from how farmers cope with crop failure due to unpredictable weather and drought. It might mean finding solutions for marine plastics by using GPS on shipping data, satellite data. It would also mean building on the knowledge of fisherfolk in addition to new sources of data. For us, a successful lab would help decision makers diversify their sources of evidence beyond the norm — census and survey data. A successful lab would also be able to build upon the knowledge of people living on the margins to change how we do development.
We’re betting that diverse sources of information and inspiration from people solving their own problems will make us more real time, more responsive and closer to people living in poverty or facing the effects of climate change. Another related sign of success for our labs will be how diverse their portfolios of experiments are. We’re looking to spread our bets so conflicting, opposing hypotheses are welcome, as are those that cover the full range of regulatory, behavioral and technological solutions to address complex problems.
Decision Making: where learning meets action
We’re running an experiment at scale to create a learning network of labs. Building on our sense-making journey, one thing is clear: learning meets action when people make decisions based on new intelligence.
What could this look like? It could be mayors and ministers taking decisions based on real time data, and citizens’ knowledge and experience. Maybe it means countries in Africa having machine learning at their fingertips and the privacy and ethical frameworks to protect people from adverse effects of combining data sources. Maybe it means climate campaigners designing their outreach based on the latest behavioural insights showing us how the human mind can comprehend existential threats. Lots of possibilities. But bottom line: a sign of success for our lab network is someone somewhere making a decision or a pivot (that involves human or financial resources) based on the solutions, insights and tests coming from our labs.
Spreading the lab love through partner take up
Another sign of success for our lab network might also take the form of spreading of the model and its methods. For the UNDP Accelerator labs, we’ve seen early signs of this sprouting up already — for example in Da Nang, Vietnam, where the city will set up a lab inspired by the Accelerator Lab design. There are other potential spin offs in play in South Asia as well. Inside UNDP, we are seeing many UNDP offices beyond the original 60 interested in setting up Accelerator labs. Of course, we are not necessarily imagining that the structure of the Accelerator lab is the only way to see the approach disseminating.
The sign of success we are looking for are the ripple effects and spreading of the methods to infiltrate mainstream policy choices. So success could take the form of spreading such as:
- National innovation policies that value grassroots innovations such as those in Nepal and India
- Governments using experimentation before designing large scale reforms
- Real time data and collective intelligence methods seen as the go-to approach to undertake situation analyses
Value for Money and the Serendipitous Path — we need help on this one
As this experiment unfolds, these are early thoughts on what success looks like. Our team is working on a performance framework that vets this out more comprehensively. As we do, this is a genuine call for advice, input and ideas: how can the UNDP Accelerator lab network deliver on our promise while still following the serendipitous emergence that will unfold? If you’ve got methods you’ve used and seem promising, or if you dream of a non-linear way to show value for money — join the UNDP accelerator lab movement!
Huge thanks to Bas Leurs for the inspiration, feedback, visuals here. Not to mention the future application of your magic to take what is useful from these disconnected thoughts to inform a coherent Accelerator Lab performance framework.
