The same search in several of your Amazon campaigns: check the dates before you fix it
Open your Sponsored Products search term report, sort it by search term, and you will often find the same search in two, three or more campaigns. An auto campaign found it, a broad campaign also matches it, and later you added it as an exact keyword. The usual name for this is cannibalization, and the usual advice is to clean it up.
We measured how often this happens on real accounts, and what it costs when it does. The short answer: it happens all the time, most of it is harmless, and the part that is not harmless shows up as one search bought at two prices.
The measured numbers below come from the Sponsored Products search term report of 12 advertisers in Europe, the UK and the US, day by day, from March 28 to October 16, 2026: 44,987 daily rows and 64,315 clicks. We left out the rows where the "search term" is a product ASIN (38.2% of the clicks), so what remains are 39,767 clicks on typed searches. A search is counted inside one advertiser and one marketplace, so currencies never mix.
Most searches that several campaigns served were a handover, not a fight
- 1,434 searches got clicks in at least two campaigns during the period.
- These searches brought 69.3% of all clicks on typed searches.
- 1,044 of the 1,434 never showed in two campaigns on the same day. One campaign stopped showing for the search, and another one took over. That is what happens when you move a search from an auto campaign into its own exact campaign, or replace an old campaign.
- Of all clicks on the 1,434 searches, 28.0% came on a day when two or more campaigns showed an ad for that search. That is 19.4% of all clicks on typed searches.
So a search that appears in several campaigns is not, by itself, a problem. In this data, seven out of ten such searches were simply a timeline.
Where they ran on the same days, the same search had several prices
To compare campaigns on the same search, we only counted the days on which at least two campaigns showed an ad for it. We kept the searches where at least two campaigns had 10 or more clicks on those shared days: 30 searches in 12 advertiser-marketplace pairs. In 21 of them, three or more campaigns were involved, in one case ten.
- The most expensive campaign paid a median 1.59 times the cost per click of the cheapest one for the same search, on the same days.
- The gap was at least 25% in 23 of the 30 searches. The largest was 3.89 times.
- The higher price did not buy better buyers. In the 9 searches where the cheapest and the most expensive campaign had at least 5 orders together, the expensive one converted worse in 5 of the 9.
The pattern holds with other limits. With 5 clicks per campaign: 57 searches, median gap 1.60 times, at least 25% in 47. With 20 clicks: 21 searches, median gap 1.42 times, at least 25% in 16.
One example, a German search on one account: ten campaigns showed ads for it on the same days. Cost per click ran from 0.35 € to 1.36 €, and ACOS from 16.7% to 162%. The campaign that took the most clicks, an auto campaign, paid 0.48 € per click at 21.3% ACOS.
The exact campaign was usually not the cheap one
The common advice is that once a search proves itself, it belongs in an exact campaign, and the other campaigns should not serve it. Our numbers do not support "exact is cheaper" as a rule. In 18 of the 30 searches, the cheapest and the most expensive campaign used different targeting:
- The auto campaign was the cheaper side in 11 of them, and the more expensive side in 1.
- A pure exact campaign was the more expensive side in 6, the cheaper side in 2.
We wrote about the same effect for single keywords in exact match is not the cheap one. It also holds when the same search runs in several campaigns at once.
What we could not see
The report is a daily total per campaign and search. It does not show single auctions, so we cannot say whether two of your campaigns were in the same auction, or which one Amazon picked and why. The price difference can partly come from placement: one campaign can win more top-of-search slots, which cost more, and the search term report does not split by placement. Orders are counted for 7 days after the click, and an order can be for another product of yours than the one in the ad. The 10, 5 and 20 click limits are our choice, not a measurement.
Check it in your account in 15 minutes
- Download the Sponsored Products search term report for the last 60 days, daily.
- Leave out the rows where the search term is an ASIN (it starts with "b0").
- For each search, list the campaigns that got clicks for it, with the dates.
- If the campaigns never showed on the same days, nothing to fix: it was a handover.
- If two or more ran on the same days, compare their cost per click, conversion rate and ACOS on those days only.
- Keep the search in the campaign that is cheaper at a similar conversion rate, whatever its match type. In the others, add it as a negative exact keyword.
- Check again two weeks later: the remaining campaign gets more of the traffic, and its cost per click can change.
The same search in two campaigns is not a mistake. Paying two prices for it on the same day, without knowing it, is.