In a recent blog post that first appeared on Harvard Business Blogging Network we discuss International Development. Trying to find solutions for the problems faced by people living at the bottom of the pyramid is frustrating and hugely expensive; experts say that three out of four development projects in Africa fail. PlayPump, a children’s carousel designed to help pump clean water in Africa is a typical example. Despite getting $16 million in funding, it failed to gain traction.

But the high failure rate might actually be quite simple to fix. NGOs and governments typically manage projects like PlayPump according to a fixed plan, as if they were building a dam or some other large infrastructure project.   What if they were to apply Lean Startup’s methodology of prototyping, experimenting, and pivoting to these projects instead?

Here’s a case in point.

Two billion people in the world do not have access to electricity. At the same time, recent market based analysis shows that this represents a proven cash-based market. According to a recent IFC energy access report, poor households spend a total of $37 billion/year (10-40% of their income) on kerosene and other low-quality energy solutions. In addition to its high cost, kerosene poses serious fire and health hazards (two million people a year die from indoor pollution, more than from than malaria) and it releases millions of tons of carbon dioxide each year. Alternative technologies, such as solar, are healthier and offer greater value, yet they require one-time investments of around $10-25, which are not affordable for people living on $1-2 a day.

To help these people to get off kerosene, we have been working with Nuru Energy, a for profit social enterprise working in Rwanda, where 85% of the population lacks access to the energy grid. Nuru is dedicated to “replacing the use of expensive, unhealthy and dangerous kerosene as a source of lighting for the two billion people without access to electricity”. Seed funded by the World Bank, Nuru offers ultra-affordable portable LED light solutions through an innovative micro-payment business model. The company is currently offering the most popular off-grid lighting product in Rwanda, already accounting for 45% of monthly sales by volume in 2012.

The crucial factors in Nuru’s solution are (1) the Nuru Light, a small portable unbreakable battery-powered LED light about the size of a hockey puck and (2) a recharging technology, called the POWERCycle generator, which is essentially a stationary bicycle (see a recent case study about Nuru Energy). While the Nuru Light part of the puzzle seemed clear from the start, the company had to do a lot of quick experimentation and pivoting before nailing the POWERCycle part of the solution.

First, an AC-charging solution was scrapped, as most of the population had no grid access. A solar panel solution was attempted next. But although it could charge 5 lights simultaneously, it was both expensive and the charging cycle was long. Next they tried a hand-charging device, but it required a lot of physical effort to use. This is when the POWERCycle was proposed and successfully tested.

Of course, it is both impractical and expensive to have POWERCycles at every household so next came experiments with the distribution and financing structure. Nuru set up a network of rural entrepreneurs who were responsible for distributing Nuru lights and selling recharges through POWERCycles. Nuru initially partnered with microfinance organizations to select and finance micro-entrepreneurs, but this approach was slow at getting traction. So Nuru pivoted to a micro-franchise system of partially subsidizing equipment purchase and sharing recharge revenues 50-50 with rural entrepreneurs (selected with help from local leaders and NGOs).

The income from the Nuru micro-franchise helped get the rural micro-franchisees out of poverty, but several experiments were required to nail the correct incentives and tradeoffs between take-up and margin. For example, when take-up rates proved lower than expected, Nuru dropped the upfront price of the lamp below cost (from $6 to $1.5), making up for the margin loss in additional recharge revenues.  Payment collection costs were the next bottleneck to scaling; collecting cash was difficult and expensive, so Nuru came up with a new SMS-based technology that could remotely unlock the POWERcycle and allowed entrepreneurs to pay for charging credits using their mobile phone.

Nuru’s experiments do not stop here: we have just received grants from the International Growth Center and INSEAD to conduct further randomized experiments aimed at increasing adoption of Nuru lights. We are launching a pilot to explore various mechanisms to reduce consumer inconvenience costs and behavioral nudges to increase usage of the lights. We will experiment with different revenue models, as well as giving households a subsidy to purchase additional lights, since we learned that some households still use kerosene as a backup solution. We also plan to equip Nuru lights with electronic registers that will allow us to track exact purchase and recharging patterns per household.

Experimentation at the bottom of the pyramid economies is hard: access is difficult and the data is scarce. Yet it is impossible to understand what works and what does not without experimentation and pivoting. Rather than dispensing huge sums of money to solutions that only make sense from the top of the pyramid, we think that funding agencies can do better by allocating money to small-scale experiments and studying the results before providing further support.

This post with co-authored with Ioana Popescu, Professor of Decision Sciences at INSEAD. An expert in revenue analytics, she is currently researching new revenue models for the bottom of the pyramid, with the aim to improve access to finance, water and energy for unserved markets. She serves as a Young Global Leader for the World Economic Forum.