Which Amazon Audiences Are Worth a Higher Bid?
Amazon lets you raise your Sponsored Products bid for shoppers it picks out for you, such as people who carted your product or bought from you before. We compared 263,000 of those shoppers' clicks with everyone else's on the same campaigns and days to see which ones really buy more often.
A Sponsored Products bid is the same for every shopper by default. Since April 2025, Amazon lets you raise it for some shoppers only. You pick an audience, set a percentage, and your bid goes up by that much whenever your ad competes for a shopper in that audience.
You don't have to build these audiences. Amazon builds them and offers them to every Sponsored Products advertiser. Three of them matter here:
- Clicked or added to cart. Shoppers who clicked your product or put it in their cart, but didn't buy.
- Purchased. Shoppers who bought from your brand before.
- High interest. Shoppers whose shopping history, by Amazon's reading, shows a high interest in your product.
A boost is only worth it if those shoppers buy more often than the rest. If they don't, you pay more for the same click. So we went looking for how much more often each one buys.
Why most sellers would bet on the cart
Ask a seller which shopper is closest to buying, and the usual answer is the one who already put the product in the cart. Most remarketing advice says the same. The shopper knows the product, wanted it enough to cart it, and only needs a nudge.
High interest sounds like the weakest of the three. Amazon says little about what activity it looks at or how far back. It reads like a label you could stick on almost anyone.
How we compared them
Amazon reports each boosted audience on its own, per campaign and per day: its clicks, its orders, its spend. The campaign's own report holds the whole day. Take the audience away from the whole, and what's left is everyone else who clicked that campaign that day.
That gives a fair comparison. The audience and everyone else clicked the same ads, for the same products, on the same day. Each audience click is set against what everyone else converted at on that campaign that day. A campaign that sells well, or a busy sale day, can't make one audience look better than another.
We did this for every campaign that boosts one of these audiences in our own accounts: 94 days of Sponsored Products, July to October 2026, in seven marketplaces.
High interest wins, by a lot
High interest shoppers turn a click into an order 65% more often than everyone else on the same campaign. Their orders are a little bigger too, so each of their clicks brings in about 75% more sales.
It isn't a quirk of a few campaigns. In nine campaigns out of ten with enough sales to judge alone, High interest converts better than everyone else. Every one of our larger accounts sees it, from about a third more often at the low end to nearly three times as often at the top.
The twist: cart abandoners barely beat everyone else
The audience everyone would bet on comes last. Shoppers who clicked or carted and didn't buy convert only about 10% better than everyone else. That's real, and most accounts see it, but it's small.
There's a plausible reason. These are shoppers who looked at the product, maybe more than once, and walked away. Some found it too expensive, some bought elsewhere, some lost interest. Walking away is part of what puts a shopper in this audience.
Past buyers look strong, but the evidence is thin
Shoppers who bought from the brand before convert 64% better than everyone else, as much as High interest. Repeat buyers are a good bet for anything people buy again, and that's what this looks like.
But very few campaigns boost this audience yet, and almost all of the clicks we have come from one brand's campaigns. Within that brand it's steady: 16 of its 17 campaigns with enough sales show it. Across brands, we can't say yet.
It isn't the bigger bid doing the work
There's an obvious objection. A boost raises your bid, and a higher bid can win better spots, like the top of the search results, where shoppers can convert better. Maybe the audience isn't special and the boost just bought a better spot.
Amazon's audience report doesn't split by placement, so we can't look directly. But accounts that boost High interest a little and accounts that boost it a lot see the same lift: +65% in both groups. If the bigger bid were doing the work, bigger boosts would show bigger lifts. They don't.
The other usual checks hold up too:
- One big account? Leave out the largest one and High interest still converts 61% better.
- One lucky stretch? The first half of the period gives +67%, the second half +63%.
- Clicks that cost more? Boosted clicks do cost more. On High interest they cost about 36% more than everyone else's while bringing in 75% more sales. A dollar spent on High interest shoppers returns about 29% more sales than a dollar spent on everyone else, at the boosts sellers use today.
What about Sponsored Brands?
Sponsored Brands has its own version of these audiences, and we keep it apart. Brands ads give credit for a purchase within two weeks of a view as well as a click, so their numbers don't mix with Sponsored Products'.
Two things stand out:
- Cart abandoners do much better on Sponsored Brands, about 70% better than everyone else. It holds when we count only sales after a click, so view credit doesn't explain it. We don't know why the same shoppers respond so differently to a brand ad.
- New-to-brand shoppers convert no better than everyone else. That audience is worth reaching if growing your customer base is the goal, but it doesn't buy more often, so don't boost it expecting a higher conversion rate.
What to do
- Boost High interest, and probably by more than you do. Each High interest click brings in about 75% more sales than everyone else's. A boost of 50 to 75% matches that. Most accounts boost less, and on average their boosted clicks still return more per dollar than the rest.
- Keep cart boosts small on Sponsored Products. These clicks are worth about 10 to 20% more, so a boost much bigger than that pays more for a click than it brings back.
- Try past buyers on products people buy again. Start near the High interest boost on a few campaigns, and compare their sales per dollar with campaigns that don't boost.
- Judge every boost against the same campaign. Comparing an audience with your account's average flatters it. The fair test is the shoppers it competes with on the same ads and days.
- Read the cart audience separately on Sponsored Brands. What holds there doesn't hold on Sponsored Products.
How we tested
The data. Amazon's audience reports and campaign reports from our own accounts, for every campaign that boosts one of these audiences: July 7 to October 8, 2026 for Sponsored Products, mid-August to October 8 for Sponsored Brands, in seven marketplaces. That's 263,000 Sponsored Products clicks and 46,000 orders from the three audiences.
The same campaign, the same day. For each campaign and day, everyone else is the campaign's total less the audience's own row. The audience's clicks are set against what everyone else converted at that day, and the gaps are added up over every campaign and day.
The checks. Ranges are 90% intervals, from resampling whole campaigns and, separately, whole accounts. We left out the largest account, split the period in half, and compared accounts that boost a little with accounts that boost a lot. Orders are counted within 14 days of a click on Sponsored Products.
The limits. This is real-world data, not an experiment. Sellers choose which campaigns to boost, and they boost where they expect it to work, so a campaign nobody would boost may do worse. Amazon's audience report covers only the audiences a campaign boosts. A shopper can sit in more than one audience. Days when the audience took every click have no one to compare with, so they're left out. Amazon launched this audience as "High likelihood of purchase based on recent shopping activity", and High interest looks like the same audience under a new name, though Amazon hasn't said so. Its launch note reports a median 77% lift in return on ad spend across 1,700 campaigns. That's close to the 75% more sales per click we see, but once the higher cost of boosted clicks is counted, we see about 29% more sales per ad dollar, so Amazon's figure may not count that cost.
Test it yourself.
- Pick campaigns with steady daily sales that don't boost any audience yet.
- Split them in two at random.
- On one half, boost High interest by 50%. Leave the other half alone.
- Run for four weeks, away from Prime Day and the holiday season, and change nothing else.
- Compare sales per dollar of ad spend between the two halves, and read Amazon's audience report for each boosted campaign.
Questions or feedback? Reach out to us at [email protected].
