Let’s not Reinvent the Flat Tire: Thoughts on Poverty, Adaptation and Scaling-up


This page in:

I’ve been thinking about scale a lot lately and I’m increasingly skeptical that it is the right way to get greater development impact. Let me explain. On the one hand, how can you not ‘think big’ when there are still nearly 2.4 billion people living on less than $2 per day? But is a cookie-cutter solution to eliminating poverty feasible given the complex factors that cause poverty and the vested political interests that sustain it?

In my view, we need a more nuanced way to think about scale and a more sophisticated understanding of how ideas and innovations spread. The ‘just scale-it’ prophets argue that there are plenty of examples of successful scale. Just look at McDonalds and Starbucks or microfinance and mobile telephony in the developing world, they say. These companies grew big (though not always fast) and reached billions of customers across multiple continents despite social, economic and political differences. But are these examples of ‘unbridled success’ straightforward and did ‘solutions’ just replicate themselves from one region to another?

As with most things, the story is more complex. When I lived in the Middle East for example, more than two decades after the founding of the Grameen Bank in Bangladesh, Syrians were skeptical of micro-loans given their exorbitant ‘interest rates’. Although the idea of microfinance emerged from another Muslim country, Syrians did not understand why helping the poor required charging interest rates that were more than triple what commercial banks were charging. And so the Aga Khan Foundation, where I worked at the time, had to painstakingly explain that profits were not distributed to shareholders and revenues were re-invested to expand the reach of the program. It eventually worked but it required a great deal of sensitivity to local context and adapting solutions to local needs.

But what about global brands like McDonalds, Ikea, and Home Depot? They regularly traverse national boundaries with great success. What do they do and what factors lead to their success or failure? The story of Home Depot in China is instructive. Huggy Rao and Bob Sutton in their book Scaling Excellence explain that Home Depot’s Do-It-Yourself (DYI) ethos did not translate well into China’s Do-It-For-Me customer base. At their own peril, they ignored such differences and forged ahead opening store after store. Eventually, they realized their “American model” did not translate well in China and closed down their operations. Contrast this with Ikea’s experience in the same country. Ikea realized they were dealing with a very different cultural ethos and developed several uniquely Chinese characteristics in their strategy - easy and inexpensive home assembly and home delivery. This simple, but important adaptation led to Ikea’s successful expansion in China. Similarly, McDonald’s has learned this important lesson many times. That’s why they don’t serve their signature beef hamburgers in India. Instead, they have developed a uniquely Indian menu. So, successful scaling requires adaptation rather than simple replication, quickly learning from failure, and understanding the peculiarities of cultural context.

A mother in Africe holds up her child. Solutions to vexing and complex development challenges require a similar appreciation of context and listening to those closest to problems. It’s often the people who know their context best that pioneer the most innovative and scalable solutions. 
 I remember a trip to Burundi with UNICEF last year. They were supporting a program called Mama Lumiere based on a ‘local innovation’ discovered when children living in the same village experienced highly variable rates of malnutrition. What were their mothers doing differently? After some investigation, local health workers learned that mothers of healthier children - the positive deviants - had three notable characteristics.
For starters, they had more years of schooling than their neighbors. Second, they insisted on hand washing before meals; and third, they mixed ingredients differently when cooking (although the ingredients were the same). It was that simple. They lived in the same place and grew the same crops but one group avoided childhood stunting. The adoption of the new practice amongst mothers, however, was not simple even in the same village. Model mothers (mama lumiere) showed their neighbors what they were doing differently to bring about behavior change. But it required time and patience for women and only took hold when mothers saw differences in their children’s health.

Examination of the many attempts to eradicate poverty reveals similar lessons. Scaling is not simply about replicating ‘what works’ in one context and exporting it to another. The search for global solutions to global problems is misguided. Global problems mostly require local solutions. Successful innovations take time to incubate, display openness to learning from failure, and seldom underestimate the importance of context and adaptation. It’s a simple lesson, but one that we consistently fail to remember. In our eagerness to solve the hardest problems in the world, it’s worth paying attention to what doesn’t work as much as what does.  


Aleem Walji

Director, Innovation Labs

Join the Conversation

Johannes Linn
October 23, 2014

Of course, you are right, Aleem. It's a complex world out there. But those of us who have been advocating more systematic attention to scaling up what works are well aware of the pitfalls that you refer to; the need to test, learn and adapt; the inevitable quest to find the right drivers and create the right spaces for good things to grow; etc. But all too often the individual and institutional incentives are in favor of moving from one development project, intervention or innovation to the next without adequate follow up, learning and -- yes -- where appropriate scaling up (as so well reflected in the weak sustainability record of many World Bank projects). Of course, as World Bank experience also shows, scaling up can be done badly (or what we call "Type II error" in scaling up). But let's not dismiss scaling up. It would be great if the Bank would take a serious look at what it means to pursue a scaling up agenda effectively.

Michael Cunningham
October 04, 2014

Commonsense, a pity that it seems to be presented as a breakthrough. Travelling overland from London to India in 1972, I first realised how the aggregates with which economists trend to deal arose from myriads of personal and local actions in a great variety of circumstances. I carried that insight into policy advice.

Rafique Keshavjee
October 04, 2014

This is lucid, practical and important.
I would add that
a) enabling carefully-selected local people to help think through the application of ideas (helps with local ownership)
b) supporting peer-to-peer models of review and testing (this enhances ownership and can lead to rapid local scaling up);
c) encouraging learning early from failures by removing fear of failure from experiments
These strategies are not market-testing, but rather localized proving of concept.
I'd like to know whether the approaches have been applied with success by others. I do know that they had, over time, in Central Asia, dramatic success.