Ten to thirty percent

Run this arithmetic on a book of business and the unprofitable share usually lands somewhere between a tenth and a third. Treat that as a range to expect rather than a figure to cite, because nobody has measured it across businesses in general. Finding them is arithmetic. Deciding what to do is where operations either earns its keep or flinches.

Five steps, in order.

One, establish the real margin per account including cost to serve, using rough allocation rather than waiting for perfect data.

Two, sort and look at the bottom decile. Do not act yet.

Three, for each, determine why. Priced wrong at the start, scope crept, cost to serve grew, or the customer changed. The remedy differs completely by cause and applying one remedy to all four is how these programs fail.

Four, apply the two overrides before deciding anything. Speed to cash: an account with poor margin that pays in seven days may be worth more than a better one that pays in ninety, because cash timing is not visible in margin. Volume weighting: an account that carries fixed cost you would otherwise carry alone is doing work the margin line does not show.

Five, act, in this order of preference: reprice, rescope, re-serve at lower cost, and only then exit. Exit is last because it is irreversible and because the three cheaper remedies resolve most cases.

The step operations skips is three, because sorting and acting feels decisive. Acting without cause analysis reprices customers who were fine and exits customers who were about to grow.