Why Repeat Buyers Skew Amazon Display Campaign Results

Last Updated: October 26, 2026

By Justin W. Boggs, Chief Revenue Officer

Why a Clean ACoS Number Can Still Mean a Stalled Amazon Channel

A clean ACoS number answers a narrow question, and it never asks whether the brand gained a single new buyer. Asking that wider question belongs in Amazon Ads Management whenever display is on the table. Display campaigns often get judged on numbers like that, which look great in a report and hide the real flaw.

Click-based measurement treats a click as proof an ad worked. An NBER working paper notes that clicks have become the leading example of the new metrics in advertising science.

Being common doesn’t make a click proof that an ad found a new buyer. For how to judge the spend itself, see when Amazon display ads pay off.

The Problem With Treating Every Ad Click the Same Way

The core issue is mechanical, not strategic. Your most loyal, repeat customers are often the ones who see and click on your display ads, making performance seem much better than it is.

That builds a feedback loop where ad spend gets credit for sales that would likely have happened anyway, and weak new customer acquisition stays hidden. Measuring true customer acquisition from display campaigns means looking past the click entirely.

None of this makes repeat buyers bad customers. Selling to them again is real revenue.

The problem is the label: retention spend reported as if it were acquisition. Keep the sale, and fix how it’s counted.

How Amazon’s Attribution Window Favors Ads Closest to the Sale

A display report counts a purchase as ad-driven if it falls inside a set timeframe after the ad, with no test for whether the shopper needed the ad at all.

Lower-funnel ad products, the ones closest to an existing purchase intent, show most of their effect inside that window. Upper and middle-funnel products, the ones actually built to reach new shoppers, show much less of theirs inside the same clock.

An Amazon Science study found that within the two-week attribution window, upper- and middle-funnel ad products show only 30–50% of their total effects, compared with 60–90% for lower-funnel products.

Repeat buyers are usually the shoppers closest to buying, so their purchases land inside that window and get counted. New shoppers reached by display often buy later, after the window closes. So the same short window that counts loyal buyers’ purchases misses much of the new demand display creates.

What Happens Inside an Account That Never Separates New Demand From Repeat Demand

An account that never splits new demand from repeat demand doesn’t announce the problem. Nothing breaks, no alert fires, and the reports keep looking fine.

Budget keeps flowing to the campaigns that look best, and those are the ones selling to loyal buyers. Prospecting campaigns doing the harder work look worse next to them and get cut. A few cycles later, the account is spending more to reach the same customers it already had.

That quiet is the problem. A missing alarm is not proof of a healthy channel.

A 2024 Journal of Marketing Analytics study measured Amazon shoppers’ brand awareness, consideration and purchases for 122,000 brands and tested how they change with ad and retail actions. Awareness, consideration and purchase are separate stages, and a click report blends them into one number.

The System Amazon Already Built to Separate New Customers From Repeat Ones

Amazon didn’t leave brands guessing on this one. Its advertising console already has a report built for this split, called new-to-brand.

What New-to-Brand Reporting Actually Measures

New-to-brand metrics flag each order as either a first purchase from that shopper within the lookback period or a repeat one, then roll those flags into the same report that already shows clicks. The campaign’s new-to-brand share is the number to watch. A low share means most of its buyers already knew the brand.

Read it campaign by campaign, not across the whole account. An account-wide share blends prospecting and retargeting, which hides exactly the split you’re after.

Where Amazon Marketing Cloud Fits Above Standard Campaign Reports

Standard campaign reports stop at the metrics built into the advertising console. Amazon Marketing Cloud sits above that layer, letting a brand query signals across campaigns and touchpoints instead of reading one report at a time.

Here’s why that matters: new-to-brand data can show weak acquisition, but it can’t explain it. Querying AMC shows which ads a new shopper passed before buying, so a weak new-to-brand share can be checked against what the standard report left uncredited.

Segmenting Amazon DSP Audiences So Prospecting Spend Stops Subsidizing Retention

New-to-brand reporting tells a brand what already happened. Audience segmentation inside Amazon DSP is how a brand decides what happens next.

Prospecting and retargeting pull from two different audience pools. Treating them as one pool blends both into every report. Split them when the audience is built, not when the report is read, and the overlap never gets a chance to start.

Building a Prospecting Audience That Excludes Existing Buyers

A prospecting audience in Amazon DSP can exclude shoppers who already bought the product, using purchase-history and remarketing exclusion segments instead of broad interest or category targeting alone. That exclusion is the whole mechanism. It strips out the people most likely to click no matter what, and it leaves only the shoppers who still need a reason to buy.

Expect the prospecting campaign to look worse after the split. The easy sales now sit in the retargeting lane, so its ACoS will usually be higher than the old blended number. That’s the honest number, not a sign the campaign got weaker.

Audience exclusion doesn’t create demand. It removes the noise that hides whether the fundamentals are working: advertising investment, catalog health and brand authority. A campaign can’t acquire customers faster than the brand gives shoppers a reason to convert.

Routing Retargeting Spend Into Its Own Measured Lane

Retargeting deserves its own lane. Campaigns aimed at past purchasers or cart abandoners need their own budget line and their own ACoS benchmark, kept well away from the prospecting report. Their job is to close stalled purchases, so judge them on that.

Judge prospecting on its new-to-brand share instead. Once each lane is measured against its own job, the report shows which lane is actually finding new buyers.

Frequently Asked Questions

These questions come up every time display reporting gets a hard look. Here are straight answers, built on the mechanics.

How can I tell if my Amazon Display ads are actually winning new customers?

Check the share of orders flagged new-to-brand for each display campaign. If most buyers already knew the brand, the campaign is mostly selling to them again. To prove the ad caused new purchases, run a holdout test that keeps a matched group from seeing it.

What are Amazon’s New-to-Brand metrics and how do they work?

Each order gets flagged as a first purchase or a repeat one, based on whether that shopper bought the brand during the prior lookback period. The flags show up in the advertising console next to the campaign’s other results.

If my ACoS looks good, does that mean my display campaigns are successful?

Not necessarily. ACoS only measures spend against sales, so it cannot tell a retained buyer from a newly acquired one. A display campaign can hit every ACoS target while adding almost no new customers.

Why are repeat buyers clicking on ads for products they already own?

They’re the shoppers most likely to notice a familiar product and act on it, so they click more readily than someone who has never heard of the brand. A familiar product needs no convincing, so the ad only has to remind them it’s there.

What is TACoS and why is it a better gauge of overall account health than ACoS?

Total Advertising Cost of Sale measures ad spend against total revenue, not just ad-attributed revenue. It shows whether the whole channel is growing or just holding steady. It doesn’t split new buyers from repeat ones on its own.

Can Amazon Display campaigns fully exclude existing customers?

Not fully. In Amazon DSP, purchase-history exclusions remove the known buyers the data can match, but shared devices, gifted purchases, and gaps in purchase history keep some overlap alive. The goal is shrinking that overlap, not erasing it.

What This Means

A display campaign report is a mirror before it’s a measurement. Repeat buyers polish it, so it shows a brand exactly what it wants to see.

Pull this month’s new-to-brand share campaign by campaign and mark which ones are really retargeting. Then give prospecting and retargeting their own budgets before the next spend change, and watch TACoS through the switch.

Results vary by brand, category, and investment level.