Following our recent piece on pilots, we thought we’d share a real life example of how a ‘failed’ pilot can actually be useful — and why that’s often the only credible path to scale.
In rural Zambia in the early 2010s, one in eight children was dying of diarrhoea. The treatment — oral rehydration salts (ORS) and zinc — had been known for three decades and cost almost nothing. Yet no children who needed it were getting it. Meanwhile, Coca-Cola reliably reached the same remote villages in small retail shops where lifesaving medicines couldn’t.
That gap wasn’t just a curiosity; it pointed to something more fundamental. The public health system was the intended delivery channel and primary health care was free to the end user. In theory. In practice, long distances to clinics, stockouts and waiting times meant “free” treatment was often inaccessible when it mattered. Pharmacies were scarce and urban. Outside of that, there was effectively no distribution.
And even where supply existed, it wasn’t designed for the reality on the ground. ORS and zinc weren’t being manufactured domestically, packaging was built for clinics rather than households and no one in the system had much reason to fix that. At best, the public system could treat around 3 million cases a year. But there were nearly 8 million episodes a year, a gap that wasn’t marginal but structural. Which raises a more useful question: if clinics weren’t reaching people, who was? And who, in that system, might have a reason to do something differently?
This gap was the starting point for ColaLife, an organisation that students of MSD would do well to learn more from. Not because they followed the guidelines to the T and everything went to plan. Because it didn’t — and the team had the discipline to follow what the evidence showed rather than defend what they’d originally proposed.
They were also unusually clear, from the outset, about what they were not trying to do:
“Our vision was to improve access to ORS and zinc through improved access to a better designed product. It was not to become a manufacturer of that product, a buyer or distributor of that product, a social marketing agency, or a health training institute.”
That framing does a lot of heavy lifting. The aim wasn’t to build a parallel system that worked while the project was there and disappeared when it left. It was to test whether something could exist in the system without them — and whether anyone else would bother if they stepped back.
The original hypothesis was elegant: work with a domestic manufacturer to package ORS, zinc and a small bar of soap into a kit designed to fit in the spaces between Coca-Cola bottles, then piggyback it on Coke’s distribution network to reach rural retailers. It even won a design award. Which, in development, should usually set off alarm bells.
By most measures, the pilot worked. The manufacturer produced kits, selling 25,000 “Kit Yamoyos” – the local name of the co-packed solution – through commercial retail networks, leading to a 45% increase in ORS availability. Average distance to access ORS fell from 7km to 2km.
But whatever was driving those results, it wasn’t Coke. The pilot also showed that only 5% of kits were sold through the Coca-Cola distribution network. The rest moved through general distribution channels — the same networks that already carried soap, cooking oil and airtime vouchers.
Retailers didn’t need Coke. They just needed a product worth stocking. The distribution system that was supposed to be the innovation turned out to be largely irrelevant. What really drove uptake was product-market fit: a well-designed, affordable, non-cold-chain item that retailers could sell on their existing terms.
Post-pilot, two actors drove scale that the original design hadn’t anticipated. The Ministry of Health, persuaded by the clinical trial data, updated regulations to formally recognise the co-packaged solution — removing the need for future country-level trials. That regulatory change turned the MoH into a large institutional buyer and opened the door to procurement at volume.
The second unexpected scale actor was Shoprite, Zambia’s largest supermarket chain. Shoprite didn’t just stock Kit Yamoyo — it functioned as a de facto wholesaler for small community retailers near its branches, effectively extending the product’s reach far beyond its own store footprint. This wasn’t planned. It emerged because the product made commercial sense.
By 2019, ORS/zinc coverage in rural Zambia had increased from 0% to 34%. Since donor funding ended in 2018, 1.7 million kits have been sold without subsidy. Total donor spend over the decade: USD2.6 million, a figure that compares favourably to most health programmes with a fraction of the reach.
ColaLife is often told as a story about creative distribution. It’s actually a story about being wrong in useful ways. The Coca-Cola distribution hypothesis was wrong but testing it produced the product redesign that made everything else work. The award-winning packaging was wrong, but iterating on it post-pilot led to domestically produced zinc and soap — both had initially been imported — that made the unit economics viable at scale.
What made this work wasn’t the original idea. It was the willingness to treat the pilot as a genuine question rather than a proof of concept. The team published findings that were uncomfortable. The marquee Coke partnership had barely mattered, the award-winning design was largely irrelevant. They then built the next phase around what the evidence showed rather than what the pitch deck had promised. That kind of honesty is rarer than it should be.
One design award, several wrong hypotheses and USD2.6 million later: that’s what useful failure looks like. In this case, it also continues to save thousands of lives a year by making markets work better for the poo-er.