The claim you’ve read a hundred times is that Amazon’s “A10 algorithm” weights external sales at roughly three times the value of an internal sale, so 33 orders from Google Ads outrank 100 from Sponsored Products.
I went looking for the source. There isn’t one.
Amazon has never published an external-traffic coefficient. It has never published ranking weights for anything. And “A10” is not an Amazon system — it’s a name the seller-education market invented for a set of observed behaviour changes.
That doesn’t make external traffic worthless. It makes the reason it works different from the reason you’ve been told, and the difference matters because it changes what you should measure.
What Amazon has actually said
There is exactly one relevant published statement, and it’s carefully hedged. From Amazon Ads’ own Amazon Attribution guide, listing what non-Amazon campaigns are useful for:
“Increasing glance views of product pages, which can help support organic rankings and add to the audiences you can reach with self-service display ads”
Note what Amazon names as the mechanism: glance views. Not external traffic as a signal in its own right. And the verb is “can help support”, not “improves”.
A second, stronger-sounding line appears in Amazon’s FBA guide:
“Amazon Attribution helps you understand how non-Amazon marketing channels drive discovery and sales of your products on Amazon. Keep in mind that Amazon’s algorithm uses these metrics, along with factors like pricing and fulfillment speed, to determine product visibility in search results.“
That’s as close as Amazon gets. It’s marketing-guide copy, not algorithm documentation.
And here’s the counter-evidence. Amazon publishes an official guide called “7 ways to improve your product’s search rankings.” It contains no mention of external traffic at all. If off-Amazon referrals carried a special ranking multiplier, that page is where Amazon would say so.
What Amazon’s own research paper says
Amazon has published exactly one paper on how its search ranking works: “Amazon Search: The Joy of Ranking Products”, SIGIR 2016, from A9.
It describes gradient boosted trees with pairwise objectives, trained on labels derived from purchases, add-to-basket actions and clicks — all on-Amazon behaviour. The line worth quoting:
“Behavioral features drive the rankings in Amazon Search to a much larger extent than they do in Web Search. Typically, they account for most of the variance reduction.”
External traffic is not mentioned anywhere in the paper.
So does it work, or not?
It works. The mechanism is just less exotic than the folklore.
External traffic that converts produces on-Amazon behavioural signals: glance views, clicks, add-to-baskets, purchases, and an improved conversion rate on the listing. Amazon’s own paper confirms those behavioural features dominate ranking. So off-Amazon traffic ranks you the same way any other traffic does — by generating the behaviour Amazon measures.
What has no source is the stronger claim: that Amazon detects the externality of the traffic and rewards it with a bonus.
The practical difference is significant. If you believe the multiplier story, sending traffic is the goal and volume is the metric. If you understand the actual mechanism, conversion rate is the goal, and badly-targeted external traffic that lands and bounces is worse than no traffic at all, because it drags down the conversion rate that is doing the ranking work.
That reframing is the useful part of this article. Anyone can buy clicks. The discipline is only sending traffic that buys.
The naming, cleared up
Since this comes up constantly:
| Name | Status |
|---|---|
| A9 | A real company — A9.com, founded 2003, an Amazon subsidiary that powered Amazon search. Its site shut down in 2019. Not an algorithm name. An Amazon engineer is on record: “There is no single A9 algorithm.” A9.com’s own archived site said “our ranking algorithms“, plural. The name is a numeronym: A plus nine letters spells “Algorithms”. |
| A10 | Pure community coinage. No Amazon document, paper or announcement uses it. Invented as a successor to a term Amazon never used either. |
| COSMO | Real and published. A peer-reviewed Amazon paper, SIGMOD 2024. |
| Rufus | Real product name, February 2024 to May 2026. |
| Alexa for Shopping | Real, current. |
On COSMO, since it’s now being oversold the way A10 was: it’s a system that mines commonsense knowledge from shopping behaviour to build a knowledge graph, aimed at closing “the semantic gap between queries and product catalogs.” Its live A/B test on search navigation produced a +0.7% relative sales lift and +8% navigation engagement. The eye-catching +60% figures you may see quoted are macro F1 improvements on a public benchmark dataset, not live Amazon search. COSMO is not a ranking algorithm replacing A9, and claims that “COSMO replaced keywords” are extrapolation.
On Rufus: Amazon retired the name on 13 May 2026, folding it into Alexa for Shopping — built into the main search bar, showing AI overviews above search results, with side-by-side comparisons in the results page. Default for signed-in US customers, no Prime or Echo required. Amazon frames it as unification rather than retirement, and reporting suggests Rufus continues to power parts of the experience behind the scenes. Rufus had reached over 300 million customers in 2025.
UK sellers: the Brand Referral Bonus is not available to you
This is the single most common false claim UK brands act on, and it costs real planning time.
The Brand Referral Bonus is US-store only. Amazon staff, posting in the UK Seller Central forum:
“All selling partners with a Brand Representative role assigned to a registered brand in Brand Registry selling in Amazon’s US store are eligible to enroll.”
Amazon’s own BRB page says “Sellers in Amazon’s US store.” The Amazon Ads UK page: the en-gb domain — lists availability as “North America: United States.”
Several agency blogs claim UK and EU eligibility. None of them cites an Amazon source. If your plan for external traffic economics depends on clawing back a percentage of the referral fee, and you sell in the UK store, that plan doesn’t exist.
For US sellers, the mechanics: an average bonus of 10% of the sales price on qualifying non-Amazon-driven sales, 11% for clothing, and 10% up to $200 then 7% above for furniture. It’s paid as credits offsetting future referral fees, not cash, and Amazon applies a two-month waiting period before the full amount is credited, to allow for returns. Requires Brand Registry with an active or pending trademark, plus Amazon Attribution tags.
Two things that changed the maths:
A 2025 rule prohibits claiming BRB and Amazon Associates commission on the same customer click. Penalties run to termination from either or both programmes.
And in 2026, Amazon cut Associates commissions by up to 50%, with premium categories falling from around 10% to 4–5%. Rolled out in the US around 9 March 2026, with the Operating Agreement updated on 14 April removing halo commissions — only the promoted ASIN or its variants now count. Amazon never publicly announced it. If your external traffic strategy relies on affiliate and creator partners, their economics got materially worse this year.
Amazon Attribution: available in the UK, and free
The measurement tool, unlike BRB, is available to you.
Amazon Attribution is “a free, self-service advertising and analytics measurement solution that gives marketers insight into the on-Amazon impact of their marketing strategies across paid and organic non-Amazon channels.” It’s been in beta for years and is still labelled that way.
Eligibility: sellers in Brand Registry with brand representative status, vendors, and agencies. UK is available — it appears on both of Amazon’s published country lists.
Attribution window: 14 days, last touch. The conversion must happen within 14 days of the click.
Metrics: Clicks, Detail Page Views and DPV rate, Add-to-Cart and ATC rate, Purchases and purchase rate, Units Sold, Product Sales, New-to-Brand, plus brand halo metrics.
That set is exactly what you need to test the mechanism described above. DPV rate tells you whether the traffic arrived. ATC rate and purchase rate tell you whether it was any good. If purchase rate on your external traffic is below your organic conversion rate, you are actively harming the signal you’re trying to build.
Don’t confuse this with a separate change: from 1 January 2026, Amazon changed view-based attribution for on-Amazon Store ads to a shopping-signal enhanced last-touch model with a shorter window. That affects Sponsored Brands, Sponsored Display vCPM and DSP Store inventory. Click-based attribution is unchanged, and Amazon Attribution is a different system entirely.
How to run external traffic properly
Tag everything with Amazon Attribution. Untagged external traffic is unmeasurable, which means you cannot tell good from bad, which means you’ll keep buying both.
Judge it on purchase rate, not clicks. Compare the conversion rate of each external source against your listing’s organic conversion rate. Sources below it are diluting the behavioural signal.
Send traffic to a listing that’s ready. Sending paid traffic to a listing with a weak image stack or thin reviews converts the campaign into an expensive way of lowering your own conversion rate.
Match the intent. Someone clicking a Google Shopping ad for your exact product converts. Someone clicking a broad interest ad on Meta usually doesn’t, and at scale that gap shows up in your ranking rather than just your ad account.
Don’t budget UK plans around BRB. See above.
FAQ
Does external traffic help Amazon organic ranking? Indirectly and plausibly, yes — through the on-Amazon behaviour it generates, since Amazon’s own research paper states behavioural features dominate its ranking. Amazon has never published a statement that off-Amazon referral traffic is itself a ranking input, and no coefficient or multiplier exists.
Is there a 3x multiplier for external traffic on Amazon? No. Amazon has never published ranking weights for any signal. The figure circulates widely without a source.
What is the A10 algorithm? Seller-community shorthand, not an Amazon system. Amazon has never used the name. Even “A9” refers to A9.com, a former Amazon subsidiary, rather than a single algorithm — an Amazon engineer is on record saying “there is no single A9 algorithm.”
Can UK sellers get the Brand Referral Bonus? No. Amazon states BRB is for sellers in Amazon’s US store. Agency articles claiming UK or EU eligibility cite no Amazon source.
Is Amazon Attribution available in the UK? Yes, and it’s free. It requires Brand Registry with brand representative status, and uses a 14-day last-touch attribution window.
Sami Sultan runs Ecom Enable, an Amazon growth partner for DTC brands. Programme eligibility and Amazon statements verified 27 August 2026. Where this article says Amazon has not published something, that reflects a search of Amazon’s own documentation, help pages and research publications.
Sources: Basics of Amazon Attribution; Amazon Ads FBA guide; Amazon’s own SEO guidance; Amazon Search: The Joy of Ranking Products (SIGIR ’16); COSMO (SIGMOD ’24); Brand Referral Bonus and BRB eligibility, Amazon staff, UK Seller Forums; Meet Alexa for Shopping; Wikipedia on A9.com.