What Is Amazon's Conversion Rate — And What Do You Do When Yours Is Bad?

The Two Numbers People Mean When They Say 'Amazon Conversion Rate'

There are two different metrics hiding under one name, and mixing them up is the single most common mistake in this category. Seller Central's version lives in Business Reports under Detail Page Sales and Traffic, labeled Unit Session Percentage: units ordered divided by sessions, times 100. It measures how well your listing converts a visit into an order, regardless of where that visit came from.

The advertising version — CVR inside Sponsored Products, Sponsored Brands, or DSP reporting — is purchases divided by clicks (or sometimes impressions, for view-through). It measures how well a specific ad or campaign converts a click into a sale, not how well your whole listing performs.

These two numbers will almost never match, and that's fine. A listing can have a healthy 12% Unit Session Percentage overall while a single prospecting DSP line sits at 2% CVR, because that line is showing your product to people who've never searched for it. Comparing the two directly and concluding the ad campaign is 'underperforming the listing' is comparing a session to a cold click — different denominators, different audiences.

The Formula, Worked on Real Numbers

The formula itself is simple: conversion rate = conversions ÷ audience size × 100. The audience size is sessions for listing CVR, clicks for most ad CVR. Here's what that looks like run against an actual book of business rather than a hypothetical.

Across 30 advertisers reMKTR managed in July 2026, the portfolio ran 78.4 million impressions at a $4.00 CPM — that's roughly $313,600 in media spend. Blended cost-per-click across that same book was $1.42, which puts clicks at roughly 220,000. Attributed purchases across that traffic were 57,137. Run the formula: 57,137 ÷ ~220,000 × 100 ≈ 26% click-to-purchase rate.

That number on its own looks great. It's also incomplete without two other figures from the same book: a blended cost-per-acquisition of $5.49, and 20.1% of those purchases coming from shoppers new to the brand. A high CVR built almost entirely on retargeting warm shoppers who were already going to buy will always look better than a CVR built on cold prospecting — and it will always grow the brand less. Conversion rate tells you what happened. It doesn't tell you whether your spend caused it.

What Counts as a Good Conversion Rate

There's no honest single benchmark here, and any page that hands you one number for every category is guessing. Conversion rate moves with:

• Category: consumables and low-price impulse items convert far more often than considered, high-ticket purchases.
• Traffic source: branded search traffic converts higher than non-branded; non-branded converts higher than cold display prospecting.
• Price point and review count: a listing with thin reviews or a price above its closest competitors will convert lower no matter how good the ad is.
• Season: Q4 and Prime Day traffic convert differently than a normal Tuesday in March.

The useful comparison isn't your CVR against an industry number pulled from a listicle. It's your CVR against your own trailing history, and against the closest comparable ASINs in your own catalog. That comparison actually holds still long enough to act on.

When the Number Is Bad News: What to Check Before You Touch Bids

A dropping conversion rate sends most people straight to the bidding console. That's usually the wrong first stop. Check, in this order:

• Attribution window mismatch — are you comparing a 7-day click window this week to a 14-day window last month? The rate moved because the ruler moved.
• Buy box and inventory — an out-of-stock or suppressed buy box tanks conversion silently while clicks keep flowing and get charged.
• Listing health — a changed main image, a price that's no longer competitive, or a recent run of negative reviews will move CVR before any campaign setting does.
• Traffic mix shift — if more of this month's clicks came from broad prospecting instead of branded search or retargeting, a lower blended CVR is expected, not broken.

Only after ruling those out does a bid, targeting, or creative change make sense. Fixing the wrong layer wastes a testing cycle and usually makes people distrust the metric instead of the process.

The Mistake Almost Everyone Makes — Us Included

The mistake is treating a conversion rate improvement as proof that an ad caused a sale. Last-click reporting will always credit whichever ad touched the shopper last, even when that shopper had already decided to buy. A retargeting line that only reaches people who added to cart three days ago will post a beautiful CVR and tell you almost nothing about whether your spend changed anyone's mind.

We've made this exact error — reading a strong CVR on a remarketing segment as validation, when a holdout test later showed a meaningful share of those buyers would have converted with no ad at all. That's the specific reason we reconcile Sponsored Ads and DSP together in Amazon Marketing Cloud rather than reading either platform's CVR in isolation, and why we use holdout groups and matched controls instead of trusting last-click alone. It's a slower answer than a dashboard percentage. It's also the only version of that answer that's honest.