Boosting every listing: new sales or a new fee?
A marketplace boost fee follows the marketplace's own attribution rule, so some boosted sales would have happened anyway. Boost a random half of comparable listings, then compare sell-through and net proceeds after every fee.
By Joris van Huët, Founder & CEOPublished 6 min read
When a marketplace charges a boost fee on a sale, it is applying its own attribution rule, written by the party that collects the fee. Some of those sales would have happened without the boost. The only way to know how many is to boost some listings and not others, at random.
A Depop dropshipping guide doing well on YouTube this week lists boosting every listing among its tips for more sales, next to sending offers and testing price points (the video). It is common marketplace advice, and it applies to any brand that also sells on eBay, Etsy, Depop or a similar marketplace. The measurement question is what the boost adds on top of the sales you would have made anyway.
How the fee decides what counts
eBay's help page on promoted listings is explicit about its general strategy: general fees are charged "when a buyer purchases the promoted item from a general ad that any buyer clicked on in the most recent 30 days," and the item must be promoted at the time of the click and at the time of the sale (eBay Help). Other marketplaces set their own windows and conditions, so read your marketplace's fee rules for both.
A rule like that is an attribution window, and a window counts sales that follow a click whether or not the click caused them. A shopper who had already chosen your item, and clicked the promoted placement on the way to it, still makes a fee-bearing sale. That is not a trick. It is what any window does, and it means the boost report shows fee-bearing sales, not extra ones.
Why boosted and unboosted listings are not a fair comparison
If you chose which listings to boost, you probably chose your best ones: the freshest, the best priced, the ones with the strongest photos. Compare them with the rest and you compare good listings with ordinary ones, and the boost gets the credit for the listing. That is selection bias, and it makes boosting look better than it is.
The test you can run this month
A random split of your own listings:
- List comparable items: similar product, similar price, listed around the same time.
- Pair them up and flip a coin in each pair. One gets the boost, one does not. The boosted items are your treatment group, the others your control group.
- Keep price, photos and titles fixed for both items in a pair during the test.
- After a fixed period, compare sell-through (the share of items sold), days to sell, and net proceeds after every fee, boost fees included.
Then the arithmetic, with your own numbers. Extra sales from boosting are the sales in the boosted group minus the sales in the unboosted group. Boosting pays if the net proceeds from those extra sales exceed the boost fees paid on all boosted sales, including the ones that would have sold anyway.
| Boosted group | Unboosted group | |
|---|---|---|
| Items listed | Your count | Your count |
| Items sold | Your count | Your count |
| Median days to sell | Your number | Your number |
| Boost fees paid | Your total | None |
| Net proceeds after all fees | Your total | Your total |
This is a small randomized controlled trial. It answers the question for your shop, at your prices, which no marketplace dashboard can.
Two things that make it read cleanly:
- Enough pairs. A handful of pairs can be decided by one lucky sale. Use as many comparable pairs as you have, and run it long enough for both groups to sell a meaningful number of items.
- One fee setting. Keep the boost rate the same for every boosted item during the test, so the result is about boosting, not about which rate you picked.
What to do with the answer: if the boosted group sells more but the extra proceeds do not cover the fees, try a lower boost rate in a second round and compare again. If the two groups sell about the same, the boost is mostly a fee on sales you already had, and you can drop it for items like those. The same split works for the guide's other tip, sending offers: make the offer on a random half of comparable listings and compare net proceeds, because an offer is a discount on a sale that may have come anyway.
Who writes the rule
The marketplace decides which sales carry an ad fee, and it keeps the fee. Its ads report shows boosted sales, which means every fee-bearing sale, not the extra ones. That report is accurate about fees and silent about incrementality. Nobody else is going to run the control group for you.
None of this means boosting is a bad deal. For some items it will add sales at a fee you are happy to pay. The split tells you which items, and at what rate, instead of leaving it to the marketplace's report.
If you also sell on your own store
A boost can move a buyer from your store to the marketplace, where you pay a fee on a sale you might have made directly. If you run both, read the marketplace test next to your own store's orders for the same weeks. Before you add a channel, measure the ones you have, and where to sell in 2026 walks through the same logic for new marketplaces. If Amazon is one of them, selling on Amazon and your own store covers the blind spot between the two.
Related answers
- Where Should You Sell in 2026? Measure Where You Already Do First
- Before You Add a Channel, Measure the Ones You Have
- Selling on Amazon and Your Own Store: The Attribution Blind Spot
- The Coupon and Affiliate Incrementality Audit: Stop Paying for Sales You Already Had
- Incremental sales, not correlations
Frequently asked questions
Are promoted listings worth it?
Only if they add sales you would not have made anyway, at a margin that covers the fee on every boosted sale. A random split of comparable listings answers that for your own shop.Why are boosted listings not a fair comparison with unboosted ones?
If you chose what to boost, you probably chose your best listings, so the boosted group was more likely to sell anyway. Randomize which listings get the boost, and the comparison becomes fair.How does eBay decide which sales carry a promoted listing fee?
eBay's help page says general fees are charged when a buyer purchases the promoted item from a general ad that any buyer clicked on in the most recent 30 days, and the item must be promoted at the time of the click and the sale. Other marketplaces set their own rules.How long should a boost test run?
Long enough for both groups to sell a meaningful number of items, with the same boost rate throughout. Fix the length before you start, so a lucky week cannot end the test early.
Go deeper: Causal attribution, explained.
Sixty-second versions of these ideas: Causality Engine on YouTube Shorts.
Keep reading
Terms in this article
- AttributionAttribution identifies user actions that contribute to a desired outcome and assigns value to each. It reveals which marketing touchpoints drive conversions.
- Attribution WindowAttribution Window is the defined period after a user interacts with a marketing touchpoint, during which a conversion can be credited to that ad. It sets the timeframe for assigning conversion credit.
- Control GroupControl Group is a segment of an audience intentionally not exposed to a marketing campaign, used to measure the campaign's true causal impact.
- CorrelationCorrelation is a statistical measure showing a relationship between variables; it does not imply causation.
- DashboardA dashboard is a visual display of key information required to achieve specific objectives. It consolidates data onto a single screen for quick review.
- IncrementalityIncrementality measures the true causal impact of a marketing campaign. It quantifies the additional conversions or revenue directly from that activity.
- Selection BiasSelection Bias occurs when data points selected for analysis do not represent the target population. This leads to distorted findings about marketing campaign impact.
- Treatment GroupTreatment Group is the set of users exposed to a specific marketing intervention. Comparing this group to a control group shows the intervention's causal impact.