The next innovation: scaling up
Right now innovation is one of the hottest topics in development. Even the UN is all over it :-), started by pioneering work from UNICEF, now UNDP, UNHCR, OCHA and many others are creating innovation teams, strategies and labs to help their programmes (and organizations?) become more innovative.
As this great paper from Bruce Jenks and Bruce Jones nicely elaborates, The UN, and the “development system” more broadly is at a crossroads. The current business models were designed in a different age and need to change to adapt to the emerging post-2015 development agenda, but also to the reality that change is permanent and accelerating and so we need to be more agile to keep up and stay relevant and effective.
In the past the most common way to test out new ideas and programme approaches was the “pilot programme” followed by “rigorous evaluation”. The problem with this was that a lot of pilots were created, but few were scaled up to national or global programmes. The pilots would often run for years, but with a kind of loving attention and specific starting conditions that couldn’t easily be replicated (like Millennium Development Villages on a smaller scale). The other challenge was that these projects would need to be run as designed for a couple of years before they could be evaluated (preferably against a control group or even in an RCT). This means there was little scope to adapt them based on feedback, on the ground reality and changing circumstances. A lot of these pilots were also based on theories brought from the outside rather than on local participatory design.
The new wave of innovation projects has done a lot to address these shortcomings. Projects are developed using participatory approaches such as “human centred design” so they better respond to on the ground reality. They are quickly iterated based on feedback and there is an emerging (if still challenging) culture of being willing to fail quickly and start again. There are now a proliferation of trainings, manuals, guides and checklists for those who want to give this approach a try.
But looking ahead, I see a number of challenges in making the most of this welcome development. I wonder how do we make sure that this new approach has a systemic and scalable effect on how we do our work? Here I see three related challenges
1. How do we make the culture or approaches of innovation more mainstream in our organizations so that it is a regular part of how we do our work rather than another pilot project or a high-profile, high visibility approach that organizations use to say “look we’re changing” while keeping most of their operations (and resource allocations) chugging along using the existing (and tired) model.
2. How do we scale the learning from the individual innovations themselves? While there have been a number of very successful innovations developed – how do those become adopted or adapted widely enough to generate large-scale impacts on development (rather than a significant but localized effect as is currently the case).
3. How do we innovate in the area of policy, or use innovation to inform policy making. In many countries the main game in town for external development actors is increasingly support to upstream policy work. This is inherently national and thus at scale, but innovating at a policy level and creating a space for experimenting, failing and learning is politically challenging for governments.
I don’t have all the answers to these but here are a few thoughts on each of them:
On the first a key step would be to “mainstream” innovation within the work of the organization as a part of (but not the whole of ) the programming toolkit rather than as a parallel process or set of discrete projects. This would imply developing programme guidance that allows innovation and allows but also manages the risks involved and explains when and how the approach should be used. A number of our current programming instruments such as in work planning, monitoring, budgeting management and procurement/partnership need to be revised to remove bottlenecks to innovation. But this is equally a “hearts and minds” issue where there is a need for clear messages and behaviour modeling from leadership to demonstrate that it innovation is encouraged (and to avoid where senior leaders encourage innovation in their words while their actions send a quite different message). Skill building and knowledge sharing is also key for staff to learn how to do it. In this model a team of innovation experts is essential, but their role will become more one of champion, policy setter, capacity builder and advisor than one of project manager.
On scaling up the knowledge from the innovations themselves, there are a couple of challenges. Innovations are often tailor-made to a specific context and may often not have a strong body of evidence nor “theory of change” behind them that explains the underlying method through which they work. This makes it hard to extract the generalizable learning that can be applied elsewhere. But there are a few things that can be done to mitigate this problem such as systematic data collection around innovations, and formative evaluation of the most successful to better understand how they work. Another is to share concrete examples and replicate them as “first prototype” in new contexts to see how they can be rapidly adapted to fit the next context. This can both make it easier to generate ideas and also help determine which parts are common features of similar projects in different locations and which are elements that need context specific customization. Another approach is to try similar experiments in multiple locations and see which work and which don’t to see what patterns emerge. But this also requires a more coordinated approach to innovation that the “let a thousand flowers bloom approach”. Short of this enhanced networking among innovators can help spread ideas.
Another challenge with scaling up innovation though is that the skills to manage a programme at scale are different from those required to innovate in the first place. As a project becomes more successful and matures there is less ideation and evolution and more traditional “good management” and standardization. The challenge is to organize a smooth transition between these phases so the project management style evolves as the project matures. Another idea is trying to find ways that when new ideas are tested there is already some consideration of scale from the outset. This could be in the form of constraints on the initial design that only allow designs that can function at larger scale without a high degree of customization or a high degree of technical support.
On the challenge of innovating in the area of policy – while it’s not wise to fail, even if it is fast in national policy initiatives, small-scale experiments can help identify and test possible policy actions before they become officially adopted. In particular they can help not only determine the likely effects of policy, but also different approaches to implementing it in practice. This is particularly promising in the complex areas of social policy and behaviour change for which technical fixes don’t work. But this requires carving out a space for small-scale policy experiments that are designed to help influence policy at the macro level (such as the UK behavioural insights team).
A related challenge for public sector innovation is the low tolerance for failure, even for experiments, on the part of the public who funds it. An interesting approach used in the early days of New York City’s innovation work was to tap into private and foundation funding that is more tolerant of risk but in the right circumstances might be willing to invest for social good rather than just economic gain. This also has potential to be applied in other countries, and particular in the UN system where we are looking in any case to diversify funding sources. New partnerships with the private sector could actually be a spur for increased innovation both through funding and expertise. Even with traditional funding sources it might be wise to take a portfolio approach to public spending with an explicit “set-aside” for innovative activities that follows a different set of rules for planning, budgeting and evaluation allowing more risk, but only on a small part of the overall budget.
Another challenge for policy innovation is that what works in practice is not always what is popular politically and so the policy process often takes a different direction from what the evidence from experiments would indicate. This is equally a problem with traditional “evidence based advocacy” and scaling up in general. A potential advantage for innovators is that they may be able to mobilize a constituency of supporters around a successful experiment in a way that a traditional researcher cannot.
Right now innovation is literally “the new thing” but if we are not able to come up with sound approaches to mainstreaming it and scaling up the results then it may be just another development fad rather than a new way of doing business. Then innovation won’t be as “sexy” as it is now but it will be making a more lasting difference.