The Amazon Brand Referral Bonus pays back roughly 10% of the referral fee on sales you send yourself, and that traffic feeds organic rank too. Most brands run a two week test, measure it wrong and conclude it does not work. Here is how to set it up so the answer is real.
Meta Ads
The Amazon Brand Referral Bonus pays brand registered sellers roughly 10% back on the sale price of orders driven by traffic they send from outside Amazon, and that traffic also feeds the sales velocity that drives organic rank. The theory is excellent. In practice most brands run a two week Meta test, see a return below their own store’s, and stop. That conclusion is almost always drawn from broken measurement.
Amazon wants traffic it did not have to pay for, and it says so with money. The bonus is not a marketing gesture, it is a standing offer: bring your own audience, keep a share of the fee. What decides whether you collect it is not creative quality. It is whether the plumbing was in place before the first pound went out.
What is on this page
What the Brand Referral Bonus actually is
Amazon charges you a referral fee on every sale, typically around 15% depending on category. The Brand Referral Bonus gives some of that back on sales you brought in yourself from outside the platform. The credit averages roughly 10% of the qualifying sale price and is applied against your referral fees rather than paid as cash.
Two conditions matter. You need Brand Registry, and the traffic has to be tracked through Amazon Attribution. Without the tracking tags Amazon has no way of knowing the sale came from you, so no bonus is paid and, just as importantly, no reporting exists to judge the campaign by. Attribution is not optional plumbing bolted on afterwards. It is the mechanism.
The second benefit is less visible and usually larger. External traffic that converts adds to your sales velocity, and sales velocity is one of the strongest drivers of organic position, which we set out in detail in how Amazon organic rank actually works. So a well run external campaign is buying three things at once: the sales, the fee credit, and rank you would otherwise have to buy through advertising.
The four things that break the test
We have reviewed a lot of failed external traffic tests. They fail in the same four ways, usually two or three at a time.
No attribution tags. The campaign ran, some sales happened, and nobody can prove which were which. No bonus, no data, no conclusion worth having.
Traffic landing on a search results page. This one looks harmless and does real damage. It breaks attribution and it drops the shopper onto a page showing every competitor you have.
Judging it on store ROAS. The comparison is wrong on its face. You are measuring a campaign that produces three kinds of return using a metric that captures one of them.
Stopping after two weeks. The rank effect is the larger half of the return and it lags. A fortnight measures the cost precisely and almost none of the benefit.
Attribution set up after the spend has started is not measurement. It is a post mortem. Every tag, every landing decision and every reporting window belongs in place before the campaign goes live, because none of it can be applied retrospectively.
Thinking about sending external traffic to Amazon?
We will look at your setup before you spend and tell you what is missing, free.
Where the return actually shows up
This is the part that catches even experienced marketers. The three benefits appear in three different places, at three different times.
The direct purchase value appears in your ads manager within hours. The Brand Referral Bonus appears against your referral fees, on Amazon’s schedule rather than yours. The rank effect appears weeks later, in business reports, as organic sessions and organic sales rising on the terms your product now ranks better for.
Anybody reading only the first of those three is looking at the cost in full and a third of the benefit, then making a decision. That is how genuinely profitable campaigns get switched off.
What to set up before you spend anything
Four things, in this order, all before launch.
One. Amazon Attribution tags per campaign and per creative. Per creative matters more than people expect, because it is the only way to learn which message actually sells rather than which one gets clicks.
Two. Conversions API on your own site if you also sell direct, so your platform side data is not being quietly degraded by browser restrictions while you try to compare channels.
Three. A landing decision made deliberately. Product page or storefront, chosen for a reason, tested if you are unsure. Not whatever link was to hand.
Four. A measurement window that matches your category. An impulse purchase settles in days. A £400 considered purchase does not. Set the window to how your customers actually buy, and write down the date you will judge it.
The landing page decision
Three options, and the right one depends on the campaign.
The product page. Shortest path to purchase, cleanest attribution, best for a single hero product with a clear offer. The risk is that the page shows competitor ads and related products, so your traffic can wander.
A storefront page. Fewer competing distractions, better for range campaigns and brand led creative, and it lets you build a landing experience that matches the ad. Slightly longer path to the buy button.
Search results. Never. It is worth repeating because it keeps happening: it breaks attribution and puts your competitors in front of traffic you paid for.
Whichever you choose, the ad and the destination should tell the same story. External traffic converts badly when the shopper arrives somewhere that looks nothing like what they clicked, and on a phone that mismatch is even harsher, for the reasons in your listing is a mobile listing now.
What it can look like when it works
One Meta account in our book runs at a blended 9.24 purchase ROAS across £251,714 of spend and £2,326,886 of tracked purchase value, at an average order value of £118.66. That is the direct return only, before the referral bonus and before the ranking effect that came with it. New to brand share on that account sits at 81.5%, which is the number that matters most in this context: the traffic was genuinely additional rather than a discount on customers who were coming anyway.
That last point deserves emphasis. External traffic is worth paying for when it reaches people Amazon was not already showing you to. If your campaign is mostly retargeting people who already had you in their basket, you are paying to accelerate sales you had. New to brand share tells you which of the two you are doing, and the same logic applies inside Amazon’s own tools, which is the argument in new to brand, the only DSP metric worth reporting.
External traffic is not a channel test. It is three tests running at once, and only one of them finishes in a fortnight.
Running Meta or TikTok traffic into Amazon already?
Send us the accounts. We will tell you whether the measurement is telling you the truth.
How to measure it honestly
Build the assessment before you launch, so the goalposts cannot move.
Count all three returns. Direct purchase value from Attribution, plus the bonus credited against referral fees, plus the change in organic sessions and organic sales on the products involved.
Use a fair baseline. Compare the eight weeks before to the eight weeks during, on the same products, and note anything else that changed such as a price move, a deal or a stock out.
Read new to brand share. High share means you are adding customers. Low share means you are subsidising ones you already had.
Watch conversion rate on the listing. If external traffic is poorly targeted, conversion rate falls, and that has knock on effects on your paid performance too. Traffic that does not buy is not free.
Judge at eight weeks, not two. And write the date down at the start, because the temptation to call it early is strongest at week three when the cost is visible and the benefit is not.
When external traffic is not worth it
Being direct, since it is not right for everyone.
Thin margins. If contribution is 15%, the roughly 10% bonus does not rescue a campaign that needs a strong direct return to work.
A listing that does not convert. Sending expensive external traffic to a page that fails to close is the most costly way to discover it needs work. Fix the listing first.
No brand registry. No bonus. The maths has to work on direct return alone, which is a much harder ask.
No appetite for an eight week test. If the budget or the patience will not stretch to a fair test, do not start one. A test abandoned at week three produces a wrong answer you will then believe for years.
When it does fit, though, it is one of the few tactics that improves your Amazon position and your fee bill at the same time. It just has to be measured like the three part return it actually is.
Send us the account and we will tell you if external traffic is worth testing
Free, no card, written answer in three working days, and yours to keep either way.
Frequently asked questions
It is a credit Amazon pays brand registered sellers on sales driven by traffic they send from outside Amazon. The bonus averages around 10% of the qualifying sale price and is applied against your referral fees. It only works on sales tracked through Amazon Attribution, so the tracking is not optional plumbing, it is the whole mechanism.
Yes. Without Attribution tags Amazon cannot see that the sale came from your traffic, so no bonus is paid and no reporting exists. Every campaign and ideally every creative needs its own tag, created before the spend starts rather than added afterwards.
On the product page itself, or on a storefront page built for the campaign. Never on a search results page. Search result landing looks harmless, and it damages both attribution and conversion rate at once because the shopper arrives at a page full of competitors.
Four reasons, usually together: no attribution tags, traffic landing on the wrong page, judging performance against store ROAS while ignoring the bonus and the rank effect, and stopping after two weeks. The last one is fatal on its own, because the rank benefit arrives later than the cost.
Traffic that converts contributes to sales velocity, and sales velocity is a major driver of organic position. So the effect is real but indirect: it comes from the sales, not the clicks. Sending large volumes of poorly targeted traffic that does not buy will not help and can hurt conversion rate.
At least eight weeks, and longer in considered purchase categories. Two weeks measures the cost accurately and almost none of the benefit, since the ranking effect and the bonus both show up later and in different reports.
It depends on what you need. Your own site keeps the customer data and the full margin. Amazon converts better for most products, earns the bonus and feeds rank. Many brands run both and decide per campaign rather than picking one answer forever.
It helps but it rarely rescues a campaign that does not work otherwise. Treat roughly 10% back on qualifying sales as an improvement to the maths rather than the reason to run external traffic. If the campaign only works because of the bonus, it is fragile.
Send the right reports. Get a straight answer back.
We read the account properly and write back with where the money is going, what it is buying, and the three things we would change first. You keep it either way.
No obligation · no card · a person replies, usually the same working day