Do Amazon Display Ads Secretly Boost Your Organic Sales?
Every Amazon sale counts toward your sales rank, so display ads should pay twice: once in the ad report, and again in organic sales. We followed about 20,000 products to see if they do.
Every sale on Amazon does two jobs. It pays you, and it feeds the systems that decide which products shoppers see next: Best Sellers Rank, search placement, the "customers also bought" row.
So here's a tempting idea. Display ads bring in sales. Those sales lift your rank. A better rank brings free organic sales. If that's true, display ads are worth more than their reports show, and you should bid them higher.
We tested the idea on about 20,000 products from our own accounts. The halo isn't there. Why it looks like it's there is the interesting part.
Why people believe it
The belief isn't silly. Amazon's help pages say Best Sellers Rank is built from all of a product's orders, with recent ones counting more. Nothing says ad orders are left out. Search placement follows sales history too: "As your sales of a product increase, so does your placement." And "customers also bought" comes from what people buy together, whichever ad they clicked first.
Amazon's own researchers have even measured display. Brands that added display ads on top of their other ads saw 14% higher total sales over the next 20 weeks.
That sounds like a halo. But the study counted total sales, not organic ones, and the brands chose when to start. It can't tell an extra sale from one that was coming anyway.
The catch in display's numbers
Display gets credit for a lot. A Sponsored Products ad gets credit for a purchase within a week of a click. A display ad gets two weeks, can count a view instead of a click, and can claim anything from the same brand, even a sale made by another seller.
Then there's who display reaches. Its most common use is remarketing: showing a product again to shoppers who already looked at it. Many of them were about to buy anyway. Randomized experiments on retargeting find that most of the purchases it gets credit for would have happened without the ad.
So the question isn't whether a display sale helps your rank. Any sale does. The question is whether display creates sales that wouldn't have happened otherwise.
The twist: display follows sales
At first our data seemed to back the halo. Products sold more in the weeks their display spend was high.
Then we looked at the timing. Display spend doesn't rise before sales do. It rises after.
That's partly how display works. A remarketing audience is made of people who viewed the product, so it grows whenever the product is already selling. More viewers means a bigger audience, which means more display spend. The sales come first.
Switching display on or off barely moved a product's total orders: about 3%, the same as for products where nothing changed at all in the same weeks. Sales even dipped slightly both when display was switched on and when it was switched off. A real effect would push those two in opposite directions.
Sponsored Products is different
We ran exactly the same checks on Sponsored Products. That matters. If our checks couldn't find a lift anywhere, a "no" for display wouldn't mean much.
Switching Sponsored Products on raised a product's total orders by about half. And every order Amazon credits to a Sponsored Products ad brings 0.3 to 0.5 extra orders within eight weeks: sales that never show up in the ad report.
So there is a halo, just not where people look for it. Sponsored Products sales come from searches, which is exactly what Amazon's search ranking learns from. Display mostly shows up where demand already is.
What to do
- Value display at what its report says, and no more. Don't add an organic bonus to display bids.
- Don't push display to climb the rankings. A display sale helps your rank only if it's a sale you wouldn't have made anyway, and then no more than any other sale.
- Give the organic credit to Sponsored Products. That's where the extra orders show up.
- If display is a big part of a product's sales, test it. That goes double if you use it to win new shoppers, say on competitors' product pages, rather than to remind people who already looked. The steps are below.
How we tested
The data. About 20,000 products from our own accounts, in 17 Amazon marketplaces, from January 2024 to August 2026. For each product and week we had total orders, ad spend and ad-credited orders for each ad type, Best Sellers Rank, search impressions and stock levels, so we could leave out weeks when a product was out of stock.
Why total orders. Display's reports claim so much that a product's "organic" sales shrink every time display claims one. How Amazon hands out credit can't change total orders, so that's what we judged display by.
Four checks.
- Switching display on or off. We compared each product with other products from the same account over the same weeks, and with products where nothing changed. We left out switches made alongside big changes to other ads, which was most of them.
- What comes first. Next month's spend can't cause this month's sales. If it still lines up with them, sales are driving the spend. For display it lined up almost as well as this month's spend did. For Sponsored Products it didn't.
- Best Sellers Rank. More orders meant a better rank, and it made no difference whether they came from Sponsored Products, display or organic search. A display sale counts like any other. Display just doesn't make many extra ones.
- Search visibility. Search impressions didn't rise after display started.
A second analysis, on different products with a different method, found the same thing.
The limits. Display is a small slice of ad spend in these accounts, so we can rule out a big halo but not a small one. We followed products for eight weeks, so a halo that builds over many months would escape us. And this is real-world data, not a controlled experiment: the accounts chose when to switch display.
Test it yourself. If display is a real share of a product's sales:
- Pick products where display brings in at least a tenth of the orders.
- Pair similar products and pause display on one of each pair, chosen at random.
- Change nothing else: no other ad changes, no price changes or coupons, and no Prime Day or holiday season.
- Run for eight weeks, wait two, then swap which product has display paused and run eight more.
- Compare total orders. Expect to need a few hundred products unless display is a large share of sales, because small effects hide in normal week-to-week swings.
Sources
Amazon
- Amazon Ads, "Ad campaign attribution". advertising.amazon.com
- Amazon Ads, "View Attribution Updates for Amazon Store ads", January 2026. advertising.amazon.com
- Amazon Ads, "Modeled conversions for Display". advertising.amazon.com
- Amazon Ads, remarketing audiences. advertising.amazon.com
- Amazon Ads API, reporting column definitions. advertising.amazon.com
- Seller Central, Sales Rank FAQ. sellercentral.amazon.com
- Seller Central, search placement FAQ. sellercentral.amazon.com
- Sorokina & Cantú-Paz, "Amazon Search: The Joy of Ranking Products", SIGIR 2016. amazon.science
- Dong, Momma, Wang, Gao & Sun, "Behavioral Feature Boosting via Substitute Relationships for E-commerce Search", SIGIR 2026. doi.org
- Linden, Smith & York, "Amazon.com Recommendations: Item-to-Item Collaborative Filtering", IEEE Internet Computing, 2003. doi.org
- Smith & Linden, "Two Decades of Recommender Systems at Amazon.com", IEEE Internet Computing, 2017. doi.org
- Pauwels, Schnaidt & Caddeo, "Causal impact of digital display ads on advertiser performance", EMAC 2022. Paper, amazon.science
- Qin, "Lengthen your attribution window: Which digital ads have the most long-term impact?", 2023. Paper, amazon.science
Measured studies
- Johnson, Lewis & Nubbemeyer, "Ghost Ads: Improving the Economics of Measuring Online Ad Effectiveness", Journal of Marketing Research, 2017. nber.org
- Kobayashi, Johnson & Gu, "Privacy-Enhanced versus Traditional Retargeting: Ad Effectiveness in an Industry-Wide Field Experiment", working paper. ssrn.com
- Lambrecht & Tucker, "When Does Retargeting Work? Information Specificity in Online Advertising", Journal of Marketing Research, 2013. doi.org
- Gordon, Zettelmeyer, Bhargava & Chapsky, "A Comparison of Approaches to Advertising Measurement: Evidence from Big Field Experiments at Facebook", Marketing Science, 2019. doi.org
- Blake, Nosko & Tadelis, "Consumer Heterogeneity and Paid Search Effectiveness: A Large-Scale Field Experiment", Econometrica, 2015. nber.org
- Moshary, "Does Sponsored Search Advertising Augment Organic Search? Evidence from an E-Commerce Platform", Management Science, 2025. repec.org
- Abhishek, Jerath & Sharma, "The Impact of Retail Media on Online Marketplaces: Insights from a Field Experiment". columbia.edu
- Ju, Zhao & Aral, "Advertising Spillovers in Mobile Apps: Evidence from Ad Shutoffs and Store Rankings", working paper. arxiv.org
- Oestreicher-Singer & Sundararajan, "The Visible Hand? Demand Effects of Recommendation Networks in Electronic Markets", Management Science, 2012. repec.org
- Chevalier & Goolsbee, "Measuring Prices and Price Competition Online: Amazon.com and BarnesandNoble.com", Quantitative Marketing and Economics, 2003. nber.org
Questions or feedback? Reach out to us at [email protected].
