Use contribution, not revenue
A £5 acquisition cost against a £50 order looks excellent. Against the £14 that actually remains after cost of goods, fees, shipping and returns, it is a different picture, and it is the correct one.
Return on ad spend measured against revenue is the most common way small stores convince themselves an unprofitable channel is working.
The calculation
- Take your contribution margin per order, after all variable costs
- Divide by your cost to acquire a customer
- If the result is comfortably above one, it pays on the first order
- If not, it only pays if customers come back — so check whether they do
Repeat purchase changes the answer
If a typical customer orders three times, you can afford to acquire them at up to three times the first-order contribution. That is what allows profitable advertising in categories with thin first-order margins.
But it requires actually knowing your repeat rate rather than assuming one, and many stores have never calculated it.
When advertising does not pay
- Raise average order value — bundles, thresholds, upsells
- Improve conversion, so the same spend buys more orders
- Improve retention, so each customer is worth more
- Reduce variable costs, particularly shipping and returns
- Or accept that this channel does not work at your margin
The last one is a legitimate conclusion and it is reached far too rarely.
Attribution is imperfect everywhere
| Trap | Reality |
|---|---|
| Platform-reported conversions | Platforms claim credit generously |
| Last-click only | Ignores everything that led up to it |
| Ignoring organic baseline | Some of those orders would have happened anyway |
| Short attribution windows | Misses considered purchases |
A rough measure applied consistently beats a sophisticated one nobody trusts. Watch total revenue against total spend as a sanity check.