D2C Footwear: $1.2M a Year of Ineffective Discounting, Identified and Eliminated

Background

Calderline sold footwear direct to consumer at $9.5M in trailing-twelve-month net revenue. The commercial calendar carried eleven promotional events per year at depths of 22–35%, plus an evergreen 10% code redeemed by roughly one in six non-event buyers. In total, 57% of revenue was transacted on some form of discount. Each event produced a visible revenue spike, and the calendar had become the de facto growth plan.

The situation

A promotion sells to three populations: customers genuinely won by the price cut; customers who merely bought earlier than they otherwise would have; and customers who would have bought that week at full price regardless. Only the first justifies the discount. Separating them requires a baseline demand model — an estimate of what would have sold without the event — which the company built for the first time, then applied to every event in the trailing two years.

Three findings followed. $1.22M per year of discounts went to sales that would have occurred at full price — roughly 40% of annual contribution profit. Forty percent of each event's apparent lift was pull-forward, purchases shifted from the following weeks and visible as a post-event trough. And discounted units returned at 23% against 12.5% for full-price units: sale shoppers bracket-bought sizes and returned the surplus.

The intervention

1. 01Cut the calendar from eleven events to five, each required — on projection before approval and on measurement after — to create more contribution profit from genuinely incremental sales than it surrendered in discounts to customers who would have bought anyway.
2. 02Discontinued the evergreen code.
3. 03Sized and planned the expected revenue dip from retraining deal-waiting customers, rather than discovering it at quarter end.
4. 04Introduced fit guidance, reducing sale-unit returns from 23% to 17%.
5. 05Held advertising spend flat so pricing policy was the only moving variable.
6. 06Made the deal-waiting share a monitored metric with a ceiling.

Results — twelve months following

Revenue declined by design: part of the prior growth had been borrowed timing and part margin surrendered. Realized price rose 12% without any change to the price list.

One cost proved permanent: the deal-waiting share settled at 24%, above its original 21%. A portion of the customer base remained conditioned to discounts after the calendar that trained it was gone.