Why Your Amazon Ad Data Doesn't Match Your Sales (Attribution, Explained)

The ads dashboard said 200 units. The whole account had sold 120. Their question was simple: how can ads report more units sold than the business actually sold? If you've ever looked at your Amazon ad dashboard and thought the numbers looked too good to be true, this is worth understanding, because the answer changes how you should read every report you get.

Why the dashboard looked great (and that was the problem)

On the advertising dashboard, everything looked fantastic: low ACoS, strong CTR, strong conversion rate, strong sales. The natural reaction to numbers like that is to scale, more budget, push the campaigns, they're printing money.

This brand did something smarter. They compared the ad dashboard to their actual business data, and the story stopped making sense. You can't sell 200 units through ads if the business only sold 120.

So the real question wasn't are the ads working. It was where are these extra ad sales coming from.

Step one, verify the source of truth

The first thing we do in a situation like this is check Amazon's business reports, the data that tracks actual units ordered by customers. Those numbers matched the brand's P&L perfectly. The real units sold were correct.

That told us the discrepancy was coming from the advertising side, not the business. So we isolated the campaigns tied to the ASINs showing the mismatch, and a pattern appeared fast: almost all of them were running Sponsored Display, and many specifically vCPM Sponsored Display campaigns.

Mechanism 1: the buy box traffic leak

Most advertisers overlook buy box ownership. Sponsored Products ads usually stop running when you lose the buy box. But Sponsored Display can keep serving even when you don't own the buy box on the listing.

Picture it: your ad sends a shopper to your product page, but another seller owns the buy box. The shopper buys, from that other seller, not you. You paid for the click that brought them there, and the sale still gets attributed inside your ad console.

Across hundreds of clicks, you're paying for traffic while someone else captures the sale, and your dashboard quietly counts it as a win.

Mechanism 2: view-through attribution

This one's more subtle, and it's specific to vCPM Sponsored Display. A shopper browses Amazon and sees your Sponsored Display ad but doesn't click. Later, they search again, click your Sponsored Products ad, and buy.

Now the Sponsored Products campaign gets click attribution, and the Sponsored Display campaign may also get view-through attribution, because the shopper saw that ad earlier. Two campaigns, one purchase, counted twice.

Multiply that across many campaigns and the console starts reporting sales in ways that don't match the real business numbers.

Why the numbers looked impossible

Now combine them. Some purchases happen on listings where you don't own the buy box. Others get counted twice through view-through attribution. Stack enough of both across enough campaigns, and the advertising console can genuinely report more units attributed to ads than the business actually sold.

That's exactly what this brand was seeing. Once we explained the two mechanisms, the mystery dissolved.

This isn't Sponsored Display is bad

It's the opposite. Sponsored Display is genuinely useful, it expands reach, keeps your products visible across Amazon, and influences shoppers earlier in the journey.

The lesson is about how you read it. If you interpret Sponsored Display attribution the same way you read Sponsored Products performance, you'll overestimate how much revenue those campaigns actually drive, and then scale the winners into spend that isn't really winning.

How to check your own account

A quick gut-check you can run today. Compare your ad-attributed units against your total units ordered (business reports) for the same period; if ads are claiming a suspiciously large share, dig in. Look at which ASINs lean on Sponsored Display or vCPM, those are where the gap usually hides. Check buy box ownership on your top-spend ASINs, since if you're losing it, some of that ad-driven revenue is going to another seller. Separate attribution metrics from incremental sales in your head, since the dashboard tells you what was attributed, not always what your ads actually caused.

The goal was never just to run ads. It's to understand what's actually driving the sale.

Common Questions

Why does my Amazon ad data not match my sales?

Usually attribution, not a bug. Two common causes: Sponsored Display serving on listings where you've lost the buy box (so another seller captures the sale you paid to send), and view-through attribution on vCPM campaigns counting a purchase that a different campaign also claimed. Always reconcile the ad dashboard against Amazon's business reports.

What is view-through attribution?

It's a sale credited to an ad the shopper saw but didn't click, before later buying through a different path. On Amazon it shows up with vCPM Sponsored Display, and it can cause two campaigns to claim the same purchase.

Can Sponsored Display steal my buy box sales?

Not steal, but it can keep sending paid traffic to a listing where you've lost the buy box, so the sale goes to whoever owns it while you pay for the click and still see it attributed to your ad.

Is attribution the same as incremental sales?

No. Attribution is what the platform credits to an ad. Incremental sales are the purchases that wouldn't have happened without it. Scaling on attribution alone is how brands overspend on ads that look efficient but aren't moving the business.