What does a supermarket stockout actually measure when national output is rising? Kenya's packaged milk shelf went thin through late October, and the argument about the cause has turned into a test of how shelf gaps get attributed. President William Ruto backed his economic adviser David Ndii on 4 November, saying two million more Kenyans are earning wages and 1.1 million of them work on affordable housing sites. The numbers behind that claim are less tidy than the claim itself: annual output rose to 5.5211 billion litres in 2025 from 5.3348 billion a year earlier, a gain of 3.5 percent, while formal deliveries to processors fell 13.2 percent between May and August 2026, from 88.89 million litres to 77.13 million, as of November 9, 2026(The Star).
The shelf gap is a channel number
The distinction that matters for a retail team is between production and what reaches a chiller. A litre counted at the farm gate never has to pass a processor, a distributor, a depot and a store receiving dock. Each of those steps can absorb volume without any change in national output, and each one fails differently: a processor runs short of raw milk, a distributor runs short of trucks, a store runs short of back room space. A single availability percentage cannot tell those three apart, which is why the Kenyan case reads as a demand story to one side and a supply story to the other.
Ndii's argument rests on a plausible mechanism. Construction sites now hand out milk alongside daily rations, and a workforce of that size does consume more dairy than it did a year ago. The problem is that no published figure converts workers into litres. The claim links an employment number to a shelf gap without the middle term, and the middle term is exactly what an availability report is supposed to supply(TechNews Kenya).
Output rose while formal intake fell
The Kenya National Bureau of Statistics series shows intake slipping for three consecutive months: 88.89 million litres in May, 84.44 million in June, 81.32 million in July and 77.13 million in August. That is a 13.2 percent decline across the quarter and 5.1 percent in the single month of July to August(Breaking Kenya News). The Kenya Dairy Board points to drought, feed costs and brokers who divert milk away from processors by paying farmers more, which is a supply chain explanation rather than a demand one.
Processor intake is not production
This is the hinge of the whole argument. A farmer who sells to a broker still produced milk; the processor simply never saw it. So national production can rise while packaged milk disappears from supermarket shelves, and both statements can be true at once. For a brand or a category team, the lesson is to stop treating intake as a proxy for supply, and to stop treating a shelf gap as proof of surging demand until the channel split has been checked.
Brokers move volume without moving output
A broker paying farmers above the processor gate price does not reduce production, it redirects it. That is why the intake series can fall while farm output rises, and why the Dairy Board points to collection and distribution rather than to cows. Whenever a channel has more than one buyer at the farm gate, intake becomes a measure of competitive pricing as much as of supply, and a brand reading it as a pure availability signal will mis-time its own replenishment.
Prices moved before the shelf did
Fresh packaged milk rose 6 percent in September and ultra high temperature milk 8 percent, according to the same statistics office dashboard. Price movement of that size usually appears after a volume constraint has already been absorbed somewhere upstream, and it gives a cleaner signal than anecdote because it is measured monthly on a fixed basket. A store that reports empty shelves plus a flat price is describing a distribution problem; empty shelves plus a rising price points to a genuine shortage of saleable product.
| Indicator | What it separates | Cadence |
|---|---|---|
| Formal intake to processors | Channel supply versus farm output | Monthly |
| On-shelf availability by category | Store level gaps versus national gaps | Weekly |
| Retail price by pack size | Scarcity versus cost pass through | Weekly |
| Rationing notices per store | Demand spikes versus allocation rules | Daily during a shock |
| Substitution rate to other brands | Loyalty loss versus temporary gap | Weekly |
Rationing matters more than it looks. When shops cap purchases, the shelf empties faster than sales would suggest, because the cap pulls forward demand from households that would otherwise buy later in the week. Availability data collected during rationing therefore overstates the shortage. A monitoring routine needs a flag for the practice, or the same gap gets counted twice: once as unmet demand and once as lost sales.
What to log when a stockout is claimed
The useful response to a public argument about causes is a short evidence trail. None of the steps below needs new systems; they need the same category and store data that a weekly availability review already produces, kept in a form that can be handed to someone outside the team.
- Fix the measurement unit. Record availability by SKU and pack size in a defined store panel, not as a national percentage.
- Attach a cause code at the point of the gap: upstream supply, working capital, allocation, or demand spike.
- Separate intake from output in the same weekly file so the channel split is visible without a new report.
- Log price and rationing alongside availability, because both change how a gap should be read.
- Review the panel after the event and keep the dated extract, so the next claim can be tested against it.
Kenya's gap will resolve or persist on its own schedule, and the political argument will move on before the data does. What survives is the measurement question it exposed: a stockout is a statement about a channel, not about a country, and the two only line up when intake, price and shelf data are read together(African Digest).
FAQ
Is a stockout the same as a shortage?
No. A stockout means a shopper could not buy the item at that store at that moment. A shortage means the volume available to the market fell. Output can rise while both stockouts and shortages appear in different places.
Why does processor intake matter more than farm output for retail?
Packaged milk on a supermarket shelf has passed a processor. Intake measures what entered that channel, so it tracks shelf availability far more closely than a national production figure that includes milk sold informally.
How should a brand read rationing by retailers?
As a distortion. Purchase caps pull demand forward and empty shelves faster than sales justify, so availability measured during rationing overstates the gap and should carry a flag in the dataset.
What single number best captures the risk?
Availability by SKU in a fixed store panel, with a cause code attached. A national percentage hides whether the gap sits upstream, in working capital, in allocation or in a demand spike.
Sources
- The Star, "Ruto backs Ndii's housing workers theory on supermarket milk shortages" (2026-11-04)
- African Digest, "Construction Workers Cited as Cause of Milk Shortages in Kenya" (2026-11-04)
- TechNews Kenya, "David Ndii Explains Why Milk Is Disappearing From Kenyan Supermarket Shelves" (2026-11-03)
- Breaking Kenya News, "Affordable housing construction workers causing milk shortage" (2026-11-04)
