Twaweza’s Varja Lipovsek, (Learning, Monitoring & Evaluation Manager) and Rakesh Rajani (Head), respond to this week’s 
series of posts on their organization’s big rethink.
That Duncan Green dedicated three posts on Twaweza’s ‘strategic pivot’ may signal that our work and theory of change are in real trouble, but we prefer to take it as a sign that these issues are of interest to many people working on transparency, accountability and citizen-driven change. His posts follow a terrific two day evaluation meeting. Here are a few clarifications and takeaways.
Spiritual matters first. We very much believe that Twaweza’s soul remains intact: we want to contribute towards change in complex systems in East Africa, by promoting and enabling citizens to be active agents and shape their lives. Our experience over the past four years has made us question much of how we ‘do’ citizen agency, but we are not quite throwing out the baby with the bathwater.
For example, in our original approach we didn’t want to be prescriptive about citizen action; we wanted to expand choices and leave it up to people to decide, what we called an ‘open architecture’ approach to social change. Sounds good; problem is that it doesn’t work so well in practice and the evidence of successful change suggests a need for less openness and more focus. New evidence about the bandwidth that poor people have to make good decisions provides useful insights on what one can realistically expect people to do.
Moreover, we have learned that we need to better articulate what we mean by citizen action – including private v public and individual v collective. We take to heart the call from the evaluators meeting (and Duncan’s blog) to both analyze what kind of action we have been promoting, and want to promote in the future, and whether we prioritize some above others, including our stance on the desirability of voice or exit.
In essence, this is a move away from an unexplained “magic sauce” model where we feed some inputs (i.e. information) into a complex system, hope 
that the (self-selecting, undifferentiated) citizens will stir it themselves, and voila – a big outcome (such as increased citizen monitoring of services, and improved service delivery) will somehow pop out on the other end.
Precisely because the processes and systems we seek to influence are nuanced, multi-layered, and steeped in politics (from local to national to international), and precisely because we no longer believe there is a single recipe to the magic sauce, we need to do a number of things with greater clarity and thought.
Second, we need to understand the systems in which we work much better, to map them out; to do the kind of “3i” analysis to which Duncan referred (others call it political economy analysis). Part of this is also just simply doing our homework: engaging more with both the theoretical frameworks and empirical evidence from within the transparency and accountability field, but also wider such as in public health, economics and political science. We know that experiences are not automatically portable across contexts, but reading deeply can help us think sharply.
Third, we accept that our original Manichean emphasis on ‘officialdom’ vs. ‘lived reality’ (government vs. people, formal governance vs. hustling) is neither an accurate representation of reality, nor a helpful way of shaping action. Enabling citizen agency means maneuvering precisely in that space between supply and demand, between citizens and state.
However, in our East African context, confidence in engaging with the formal sectors has been eroded by years of unresponsive and corrupt systems, so much so that even when there is a genuine opportunity to engage or provide feedback, citizens often don’t do so. It’s critical for us to understand the barriers and motivators for citizens to act– but equally, we need to understand the barriers and motivators from the system/sector side, and look for opportunities where the two can connect to get things done.
Duncan’s point on taking advantage of critical junctures is well taken; and although we did not mention it during the meeting, we have been responding and engaging with topical and political issues, particularly in Tanzania, for example in relation to the crisis in education, the new phone card SIM tax, and pricing of malaria medicine.
Fourth, we must be wiser about where we think we can contribute the most, while at the same time take risks and foster innovation. This last point is important. In seeking to engage with complex systems in a complex world, we need to do two things simultaneously: keep a hard line on a handful of hypotheses (both in terms of implementation and measurement – next point), as well as be nimble in experimenting with innovative approaches.
Part of this is what we are calling the “positive deviants” lab; part is the “programming lab”. The former will be an initiative to find, understand and – when possible –replicate examples of citizen action and engagement across East Africa. The latter will be an effort by us and our implementing partners to be more nimble and experimental in identifying new directions and implementation models, setting up tighter feedback loops between recipients and implementers. As our Advisory Board member Lant Pritchett tells us, you never get it right the first time. So the point is not to design the best intervention, but to develop intelligent antenna to learn and adapt fast.
Fifth, we recognize a real tension between the desire for quality, thoughtfulness and iteration on the one hand and scale on the other. The last thing we want to do is create a set of boutique programs or our own Millennium Development Village. The East African landscape is littered with thousands of pilots that went nowhere. But we think there is a way to do things in a way that has scale built in from the beginning; ingredients include simplicity (to allow easier understanding and replication), a political economy analysis of the drivers and levers of change, and keen attention to incentives and crafting winning coalitions.
The upshot of all this is to privilege learning in the organizational DNA. Sure we are, at heart, about implementation and getting things done. But it is 
precisely because we want to get things done better that we take measurement and learning so seriously (though we take the point on balancing the two). We believe that the type of analytical thinking that is inherent in evaluation is also incredibly useful in implementation. It permeates the points made above: understanding complex environments and systems, defining better citizen agency, and articulating hypotheses of how to promote it.
So how to develop a learning posture across the organization? We agree with Duncan that if learning is boiled down to the quest for hard quantitative nuggets, we will have missed not only the big picture, but the core of the complexity we seek to understand. What we are aiming to do, particularly next year, is to set up a learning architecture which will use a variety of metrics, methods, and tools; which will build on the theory behind the implementation choices, allow us to learn quickly as we implement and to vary implementation accordingly, and to look for and capture different kind of outcomes.
In sum, these changes are not about retreating from grand ambitions; they are about assessing where we have gotten so far and shifting tactics. We feel a deep responsibility to be thoughtful about our job, to do it well – the stakes are high for us, but much higher for the people whose realities we want to improve. If we didn’t hold ourselves accountable to high implementation and measurement standards, then we truly run the risk of squandering the chance to do something really powerful. Stay tuned.
Right that’s enough Twaweza posts. Got some more on Tanzania, but might hold that over for a few days. Meanwhile, here’s one final video as I head back to the UK