The share read had been clear for two quarters: roughly 80 basis points of lost share in two large general-merchandise categories, accelerating into a third. The instinct in the room was that the loss was a pricing problem. The marketplace competitor was running everyday prices roughly five per cent below ours on a thirty-SKU watchlist; if we matched, the conventional logic ran, we’d defend the share.
The pricing committee had built a defence plan. About £14m of annualised margin at risk; the team wanted authority to redeploy roughly £8m of it into matched everyday prices on the watchlist SKUs.
What the cohort view said
We spent three days doing one thing: rebuilding the share-loss number from the customer cohort up, rather than from the SKU price comparison down. The reconstruction said something different.
Of the basket of customers we’d lost share to, only about a third had genuinely switched to the marketplace on price. The other two-thirds had switched on fulfilment: a twenty-four-hour click-and-collect window we didn’t offer, against a same-day window the competitor did. Many of the price-led switchers were already lost: their average order value with us had been declining for four quarters, and the price gap was the last straw, not the cause.
The pricing committee had a defence plan. The cohort view said the share loss they were defending wasn’t the share loss they had.
Two consequences. First, matching prices on the watchlist would defend about a third of the share at full cost — the £8m would buy roughly £2.5m of structurally defensible margin, which is a poor trade. Second, the fulfilment-led switchers were the larger group, and our fulfilment window was a fixable problem rather than an unfixable price problem.
What we actually did
Two moves, both small. The pricing committee took a smaller, sharper position on a fifteen-SKU subset where the cohort data said the elasticity was structurally favourable. That cost about £3m of margin and defended roughly £5m of contribution. The remaining authority — about £5m — was diverted into a fulfilment-window pilot in three regions, where the question was whether a same-day window would recover the fulfilment-led switchers at a margin we could live with. It did, in two of three regions, at a payback inside seven months.
The full defence would have cost roughly £8m of margin to recover about £2.5m of structurally defensible contribution. The actual move cost about £3m and recovered roughly £9m across the two interventions, net of fulfilment investment. A four-times improvement on the original plan, achieved by changing the diagnosis rather than the budget.
The lesson
The pricing committee’s instinct wasn’t wrong because pricing was the wrong instrument. It was wrong because the diagnosis was a level too high: we read the share loss as a single number and matched a single instrument to it. The cohort view said the share loss was actually two problems, of which only one was a pricing problem at all.
This is the most common mistake we see in defensive commercial work: the budget gets sized against the symptom, then split across the instruments the team is most comfortable with. The trick is to keep going one level deeper than the room wants to, until the question stops being “what should we do about the share loss?” and starts being “which customers, in which categories, switched for which reason?”. Once that decomposition exists, the right instrument usually picks itself.