What is a good conversion rate on Amazon? Why the averages mislead, and the number to use instead

Ask what a good conversion rate on Amazon is and you will find a confident answer on almost every page: 8% to 12% is average, 13% to 15% is excellent, under 8% needs work. That is a composite of several of the top pages; the bands differ from page to page, and none of the pages we read set out how its figure was measured.

That matters if you are trying to judge whether your listings are any good. A product at 7% can be beating its market. A product at 18% can be losing sales it should have won. This guide explains why the averages mislead, then sets out the two comparisons we use instead. Both come from reports already in your Seller Central account.

Why the published averages mislead

Most of the figures come from agencies and software companies. They may well be drawn from real accounts, but without a method you can’t tell which accounts, which categories, which marketplace, which dates or which measure. We don’t quote them, and our rule is not to cite any statistic whose source we can’t check.

Amazon publishes no benchmark that we could find. Its advertising guide even says “There is not one sole standard conversion rate"³, meaning there is no single definition of what counts as a conversion.

Even a carefully measured average would mislead you, for six reasons. The first three come from our experience across accounts, not from Amazon figures.

• Category. A phone charger and a sofa are different decisions. Low-price, repeat and replacement purchases convert far better than considered ones.
• Price. Within one category, the cheapest products usually convert best. A premium product converting below the category can still be the more profitable one.
• Traffic mix. Shoppers who search your brand name, come back to reorder or arrive through a deal convert far better than shoppers browsing a generic search. Run more ads on broad searches and your rate falls even if the listing is unchanged.
• Units, not orders. Amazon’s best-known conversion measure counts units. More on this below, because it is the most common reason a “conversion rate” looks too good.
• Business buyers. Business Reports splits out Amazon Business customers with their own B2B columns¹. A few bulk business orders can make an ordinary listing look exceptional.
• Variations. A parent listing is a total of its children. A strong child can hide a weak one.

The two numbers in your Business Reports

Amazon’s Business Reports give you two conversion measures. These are Amazon’s definitions, word for word¹:

The difference is in the top line of each sum. On Total Order Items, Amazon adds: “An order can have multiple order items, and each order item can have multiple units"¹. A shopper who orders six of the same product in one go adds six to units and one to order items.

So read both. When unit session percentage runs well ahead of order item session percentage, look at the B2B columns, multipacks and quantity discounts before you credit the listing.

Two more details from Amazon’s glossary decide which report you use. First, “When a child ASIN is visited, it does not also count as a visit to the parent ASIN"¹. Second, in the “Detail Page Sales and Traffic” report page views are summed “only for days when there was sales activity on the ASIN”, while in the “by Child Item” version they are summed “for the entire date range, regardless of sales activity"¹. On a product that stopped selling for a while, the first report hides the traffic you still had, and can make conversion look better than it was. Use the Child Item report.

Comparison 1: your share of clicks against your share of purchases

An average from someone else’s accounts tells you little. What you want to know is whether shoppers who see you and click on you go on to buy, compared with everyone else on the same search. Amazon gives brand owners the numbers to work it out.

Search Query Performance sits in Brand Analytics. Amazon says it “provides visibility into the performance of the top search terms associated with your branded products based on customer search behaviour. The report surfaces query volume, impressions, clicks, add-to-basket and buys"². It includes “organic and Sponsored Products originating from the search results page”, with “a maximum of 1,000 query results"². To open it you need a brand representative role in Brand Registry².

Amazon’s UK help page doesn’t define share. In the accounts we work on, the dashboard shows each search’s total and your brand’s count, and share is your count ÷ the total. That is our working, not an Amazon metric.

For each search, work out two shares and one ratio:

Above 1.0, shoppers who click on you buy more often than shoppers who click on the average result for that search. Below 1.0, they buy less often. Because every brand on the search faces the same shoppers, price points and season, the ratio removes most of what makes published averages useless.

Worked example

Invented figures for one garden hose brand over one month, on its three biggest searches.

The sums, row by row: 1,600 ÷ 20,000 = 8.0% and 200 ÷ 2,000 = 10.0%, so 10.0 ÷ 8.0 = 1.25. Then 1,800 ÷ 12,000 = 15.0% and 135 ÷ 1,500 = 9.0%, so 9.0 ÷ 15.0 = 0.60. Then 250 ÷ 5,000 = 5.0% and 20 ÷ 400 = 5.0%, so 1.00.

Across all three searches, this brand turns 355 of its 3,650 clicks into purchases: 9.7%. Hold that against a published band of 8% to 12% and you would call it average and move on. The per-search view says something more useful:

• On “expandable garden hose” the brand converts a quarter better than the market (12.5% of its clicks become purchases, against 10.0% for all clicks: 2,000 ÷ 20,000). That search rewards visibility. It is a candidate for more ad spend, if the margin per order covers the extra clicks.
• On “garden hose” it wins plenty of clicks, 15% of them, and then loses the sale. Its clicks buy at 7.5% (135 ÷ 1,800) against 12.5% for the market (1,500 ÷ 12,000). Something on the product page, or the price beside the products around it, sends shoppers elsewhere. More ad spend here would buy more of the clicks that underperform.
• On “hose reel” it converts at exactly the market rate, 8.0%.

Three cautions. Search Query Performance counts only what starts on the search results page and leaves out widgets on that page such as “Top Rated” or “New Arrivals"². Its numbers will not match your other reports, and Amazon says plainly that “we do not recommend comparing the metric names of data across different dashboards"². And on small searches a handful of purchases swings the ratio, so judge it over a month or more, on searches where you have at least a few hundred clicks.

The Query Detail view then shows you who wins each search: “the top 10 ASINs for the selected search query” compared with your own². When your ratio is below 1.0, that list tells you which product page to study. Our Brand Analytics guide covers the rest of the dashboard and how to rank searches by what a point of share is worth.

Comparison 2: each product against its own history

The second comparison needs no Brand Registry. Download Detail Page Sales and Traffic by Child Item from Business Reports, for the same weeks this year and last, and put each child product beside itself.

This is how we judge our own work. On one account we manage, order item session percentage rose from 7.39% in January to August 2025 to 9.71% in the same months of 2026. That is 2.32 points, or almost a third higher (9.71 ÷ 7.39 = 1.31). We don’t offer 9.71% as a target for anyone else. The point is that the account’s own history was the yardstick, measured on the same report, the same months and the same measure.

Do this at child level, not parent. Here is why, with invented figures for one variation family over four weeks:

At 10.0% the family looks healthy. The small grey child converts at a third of its sibling’s rate (4.0% against 12.0%). Something on its page is wrong: the main image may show the large size, the size may be unclear, or its reviews may come from a different product. The family total hides it.

So what is a good conversion rate?

Our answer, and it is our view rather than an Amazon rule: a good conversion rate is one where your share of purchases is at or above your share of clicks on the searches you depend on, and where each child product holds or improves on its own history once price and promotion changes are accounted for. A product that meets both is converting well, whatever the number. A product that meets neither needs work, even at 15%.

If you aren’t in Brand Registry, use the second test alone, and add a third: compare each product with your own median for similar products. Our listing optimisation guide uses the same test.

What moves conversion

Once you know which products and searches underperform, the causes are usually on the page. Our profit audit always checks two things here, because they are where we most often find the problem:

• What the shopper sees on the search card. The price shown, the pack size, the delivery promise and any badge. We flag a card that shows the wrong pack, a price far off the cards around it, or a slow delivery date. Shoppers decide whether to click from the card, and they arrive on the page expecting what it showed them.
• Whether the listing’s facts agree with each other. Title, bullets, spec table, description and images. Two different sizes, weights or capacities on one page cost sales, because shoppers read all of them.

Beyond that, each part of the page has its own guide:

• Title, bullets, images and search terms: our listing optimisation guide, and the free Listing Checker for Amazon’s 2026 title rules.
• A+ content: our A+ content blueprint covers how to build it and how to measure it honestly.
• Reviews: our guide to getting reviews on Amazon covers what Amazon allows.
• Testing: our experiments guide covers Amazon’s Manage Your Experiments tool: which products qualify and how many visitors a reliable result needs.
• The Featured Offer: if another seller or a pricing error takes the “Add to basket” button, conversion falls with no change to your page. Our Featured Offer guide covers the checks.

Conversion also feeds straight into advertising cost. At the same cost per click, a product converting at 12% instead of 8% needs a third fewer clicks for each sale (8 ÷ 12 = 0.67). That is why we fix conversion before we add budget.

When conversion falls

A sudden drop is usually something specific: a stock-out on one child, a lost Featured Offer, a price rise, a new competitor on your main search, or a change to the page. If your sales are down too, our guide to Amazon sales falling on last year splits the drop into traffic, conversion and price, and gives the order to check things in.

How we turn conversion gaps into pounds, and which listing checks we run, is set out on our How we measure page.