Lean Innovation in International Development

This summer I am working with Polymath Ventures on a financial services venture, having previously worked in economic research in emerging markets including Malawi and Pakistan. I came to Polymath seeking exposure to an exciting startup philosophy and a high-performing culture, and to gain on-the-ground experience in Latin America. This summer has more than met my expectations (you’ll see that I’ve drank a good amount of the Polymath Kool-Aid).  With this post I would like to compare ‘mainstream international development’ and startup philosophies, with a focus on the divergence in mindsets around the speed of innovation and adjustment.

For a comparison point, I am going to use the curriculum of the MPA/ID program at the Harvard Kennedy School (which I am currently attending) as an example of a modern yet mainstream international development mindset.

Consider the two charts below. The chart on the left is a common process taught to MPA/ID students, and involves designing policy based on economic theory, impact evaluation, and achieved learnings for the next redesign cycle. The chart on the right shows an extremely similar loop, which many readers will recognize from the seminal book The Lean Startup.  ‘Lean’ thinking certainly has a large influence on Polymath’s thinking and methodology; during my first day of work I was told to read it ASAP, and it has helped provide context for much of my work this summer.

That there is such a strong overlap between the basic innovation structures is not surprising; both philosophies operate in a world with large amounts of uncertainty. Both Lean and modern development organizations want to be efficient in their allocation of resources and eliminate wasteful programs by incorporating learnings and revising initiatives. What I do find strange, however, is that speed – a key component of a ‘Lean’ process – is completely absent from the mainstream mindset.

Before I try to explain why this is the case, let’s think through a couple of contrarian examples. Consider a research project I worked on in Pakistan.  Started in 2009, this project is still ongoing, with only preliminary findings capable of being presented. No serious policy has changed, the basic structure of the project has been only marginally altered, and it takes an entire year to make even minor changes. That this project continues after four years is not strange in many development circles; however in a startup this would be horrifyingly slow.

On the other hand, Polymath has fully embraced an alternative philosophy which encourages the speed of innovation necessary in a startup. An example of this was when a team I was on tabled an idea for a financial product that relied on professional training to motivate financial advisors.  Where the MPA/ID framework would have involved diligently researching economic literature and theory to design an optimal incentive system, taking months and many man-hours, Polymath instead focused on methods to quickly push decisions forward, embracing the uncertainty implicit in these decisions.  Rather than being tied to rigorous statistical analysis to make a decision, Polymath focused on revising or killing the idea swiftly – albeit with a lower level of analytical rigor. As a result, we were able to move forward with different ideas within days, rather than months.

To be sure, speed shouldn’t always be a primary focus. Polymath can get away with this because the stakes are comparatively low – if one of Polymath’s minimum viable products (MVP) fails, the consequence is small compared to what might happen if an initiative of the World Bank fails. As a result, we can trade off more speed for less certainty in our decisions. The thing that frustrates me, however, is that the philosophy of rapid innovation and adjustment could still exist in the mainstream development sector, with the recognition that ‘rapid’ will be substantially slower than in a startup due to a higher requisite level of certainty for decision making. So why doesn’t it?

I see two major reasons for ‘rapid’ as a quality absent in traditional development thought: the recent popularity of large scale randomized controlled trials, and a semi-paternalistic opinion that outsiders know best.

Randomized controlled trials or RCT’s are currently an in-vogue concept within the international development sphere, championed by academics like Esther Duflo and Dean Karlan. An RCT is a common practice in medical trials: two groups of people are randomly assigned to either a treatment group (receiving some new drug) or a control group (receiving a placebo or existing drug).  As the two groups are randomly assigned, on average they should display the same characteristics, and any resulting difference between the two can be causally attributed to the drug assigned. Thus, an RCT in a developing world context involves randomizing people into a treatment and a control group, with any resulting differences being logically caused by the different interventions given to the groups.[2]

RCTs present the benefit of a clean causal argument in evaluation, but this causal certainty comes at a cost. These experiments often take years to progress, and have little or no room for innovation and adjustment during the experiment.  Once the ‘treatment’ is defined, it cannot be altered or the causal logic breaks down.  Adjustments can be made after an experiment runs its entire course, but this can be painfully slow.

Another source of this lack of speed may be a general spirit of latent paternalism. The thought that highly educated people (often foreigners) know what is best helps justify the implementation of certain policies, without the innate paranoia that this policy design is wrong or missing a key point. Polymath’s practice on the other hand, is reliant on the assumption that we don’t have the answer prepackaged, and we need to adjust our products and designs as frequently as possible. In fact, much of the Polymath strategy incorporates design thinking, informed by on-the-ground user interviews and a belief that many issues in the developing world are misunderstood. Our assumption of fallibility drives constant innovation, and I hope that this philosophy begins to spread across startups, NGOs, and government agencies in the developing world.

Coming to the end of my summer with Polymath, I now fully appreciate the lean mentality, and strongly believe that international development as a field would benefit from learning from the philosophy Polymath represents. In my opinion, it should become a key component of the general international development mindset, and I would recommend anyone interested in improving life in the developing world to adapt a lean philosophy. By innovating faster and smarter, we come to a stronger understanding of the difficulties and pain people face every day, and more importantly, arrive at pioneering solutions to change the world that much faster.

[1] Reis, Eric. 2011. The Lean Startup: How Today’s Entrepreneurs Use Continuous Innovation to Create Radically Successful Businesses.
[2] For more information on RCTs in the developing world, please visit //

More from  Venture