We do not sell strategy separately from the build. The people who write the plan are the people who ship it, and the plan names one number, one date and one method before anyone opens an editor.
Your core runs on something written before 2005, the people who wrote it have retired, and every change takes a quarter.
Measured 14 March 2026. Median across 8,412 consignments, Feb–Mar 2026, against an August 2025 baseline drawn from the same yards.
Their own six-engineer platform team runs the dispatch service and carries the pager.
The books are closed out of spreadsheets, three departments keep private copies of the truth, and nobody trusts the stock figure.
Measured 6 April 2026. Working days from period lock to signed management accounts, averaged over three consecutive closes.
Finance runs the close themselves. We are not in the room for it.
Every question takes a week and comes back with two different answers depending on who ran it.
Measured 30 April 2026. Share of store-days with zero shelf stock, all 62 stores, six-week window against the same window a year earlier.
Their analytics team maintains the forecast models and the replenishment rules.
You are losing money between the cart and the confirmation screen and cannot say exactly where.
Measured 2 February 2026. Sessions reaching payment success over sessions entering checkout, 30-day window, n = 214,000.
Two of their engineers ship checkout changes without us.
A model demo impressed everyone in the room and then nothing reached the people doing the work.
Measured 21 January 2026. Time from discharge order to signed summary, median of 11,900 discharges across four sites.
Their clinical systems team owns the drafting model and its review queue.
You have decided to build in-house and the first six hires will decide whether that works.
Measured 18 June 2026. Calendar days from start date to a merged change running in production, for the 11 engineers hired after March 2026.
The team hires, onboards and reviews without us. We left in June.
Usually because the thing that is broken is a decision, an incentive or a process, and no system we build would survive contact with it.