Companies raise money,then find out nobodyowns the number.

Paid acquisition looked fine on paper — blended CAC was flat for four quarters — but the bank balance kept telling a different story. Nobody in the company could name a profitable segment.

Rebuilt unit economics bottom-up by cohort instead of by channel, then repriced around the two segments that actually retained. Shut down the self-serve tier that was quietly subsidising everyone else.

Payback fell from 19 months to 7 without cutting a dollar of spend. The operating review now opens with cohort payback instead of blended CAC.

Third generation, sixty years of trading, and a working capital cycle nobody had measured since the euro. Two banks were asking questions the family could not answer in a room.

Put a thirteen-week cash forecast in place in the first month, then went line by line through terms with the eleven counterparties who accounted for most of the float.

Cash conversion doubled inside a year. The bank conversation changed from covenant questions to a larger facility.

A sponsor sixteen months from exit, a finance team of two, and a cost base that had never been benchmarked against anything.

Rebid every inland lane against a clean baseline and rebuilt the month-end close so diligence would not find surprises.

Landed cost down 14% on the same volume, and a close that survived a buy-side quality of earnings review without a restatement.