Every offer, net of returns and failed deliveries
A voucher that brought orders which came back is shown for what it earned, not for what it sold. Returns the listing could not have prevented are counted separately.
Find out which promotions paid for themselves, after the returns
Send the history of your vouchers, price cuts and free-shipping offers with the orders, returns and failed deliveries they produced. We work out which ones left you better off than doing nothing, for which shoppers, and where fixing the listing beats a discount.
Most promotions are judged on the day they run
A voucher goes out, orders go up, and the campaign looks like it worked. Weeks later some of those orders come back, some were refused at the door, and some shoppers would have bought at full price anyway. None of that reaches the campaign report.
So the same discount keeps running for everyone, including shoppers who never needed it, while a listing that leaves out the size keeps losing orders a voucher cannot win back.
Sometimes the best promotion is none, and sometimes it is a better product page.
A promotion is worth the orders that stay bought, not the ones it creates
Every order a voucher buys has to survive delivery and the return window before it earns anything.
Voucher
Illustration- Orders
- Failed deliveries
- Returns
- Stay bought
How does the Promotion Audit work?
Three steps, on data you already have. Nothing goes on your site.
01You
Export your promotion history
For each group of shoppers you can see, such as new or repeat buyers, a region, cash on delivery, and each offer you ran, including no offer: visitors, orders, failed deliveries and returns with their reasons. Your margin, return cost and failed-delivery cost. No personal data.
02Us
We measure each offer against doing nothing
Each offer's expected contribution per shopper counts the orders kept, the returns a listing could have prevented, every other return, and failed deliveries. An offer is recommended only if it still beats doing nothing at the cautious end of its range.
03Both
Test what is not yet proven
Where the history is too thin, or the gain sits inside the noise, the answer is a test, with the group and the offer named. We do not recommend an offer nobody has measured.
What you get
Where to give nothing
For each group of shoppers, whether any offer beat doing nothing, and how many shoppers each choice covers.
A listing repair priced against a discount
Fixing the size or the compatibility on a product page is one more option, with its one-off cost spread over the shoppers it reaches, set beside the vouchers it competes with.
The price shoppers actually pay
Your price against the same product elsewhere after vouchers, cashback, coins and shipping, beside the sticker gap a price tracker would report.
Tests, where the history cannot tell
The offer worth trying, and the group to try it on, whenever an option could win but has not been measured.
What the Promotion Audit will and will not do
What this does
- Judge each past offer on contribution kept, net of returns and failed deliveries
- Say where no offer beat doing nothing, and for how many shoppers
- Price a listing repair against a discount on the same terms
- Ask for a test whenever the history is too thin to tell
What it does not do
- Promise more profit or fewer returns. Nothing has been measured on your history yet, and another shop's result is not evidence about yours
- Guess an effect. Every rate comes from your history or a test you ran
- Target individual shoppers. It works on groups you can already see
- Set your prices automatically. It recommends; you decide
The ones that decide it
What data do you need?
For each group of shoppers and each offer you ran, including periods with no offer: visitors, orders, failed deliveries, and returns with their reasons. Plus your margin per kept order, what a return costs you and what a failed delivery costs you. No names, phones or addresses.
Why compare against doing nothing?
Because some shoppers buy anyway. A voucher given to them is margin spent for nothing, and the only way to see that is to compare each offer with the same group when no offer ran.
How is this different from Return Verification?
Return Verification asks whether a listing repair moved the returns it was meant to move. The Promotion Audit asks which of your offers, including a repair, left you better off than doing nothing once returns and failed deliveries are counted.
Can I call it from my own tools?
Yes. The same audit, the per-group decision and the landed-price comparison are API calls. The developer docs have them.
Is it a forecast?
No. It reads what your offers did. Where the history cannot tell, it asks for a test instead of predicting.
Start with the promotions you already ran
One category, the offers you ran on it, and the orders and returns that followed.
Talk to us
Nothing is priced until a pilot has run on your own history.