You spend money on Facebook ads. Someone clicks. They end up on Amazon. They buy your book.
Did that sale come from your ad? You don’t actually know. Not without Amazon Attribution.
That gap between “the click” and “the purchase” used to be a black hole for indie authors. You were flying blind, looking at Meta’s engagement numbers and hoping they translated to sales somewhere on the other side. Attribution closes that hole. It’s free, it’s available to all KDP authors, and most people set it up wrong or don’t set it up at all.
Here’s how it works, and what to do with it.
What Amazon Attribution Actually Does
Attribution gives you a special tracking link. You put that link in your ads, your newsletter, your bio – anywhere off Amazon where you’re sending readers. When someone clicks it and then buys (or borrows) your book within 14 days, Amazon logs that sale and traces it back to the exact link that sent them.
That’s it. A tagged URL that closes the measurement loop.
The model is called “last-touch attribution.” If someone sees your Facebook ad on Monday, then clicks your newsletter link on Thursday, then buys on Friday – the newsletter gets the credit. It’s the last click that counts. Worth knowing, because it means social ads often don’t get “credited” for sales they genuinely influenced. But we’ll come back to that.
Setting It Up
Go to advertising.amazon.com and log in with your KDP credentials. Look for “Measurement & Reporting” in the left menu, then select “Amazon Attribution.”
Create a New Campaign. Add just the book you’re promoting – not your whole catalog. One book, one campaign. This matters because any sales of other titles within the 14-day window will also appear in your data, and you don’t want that noise mixed with your controlled experiment.
Inside that campaign, create Ad Groups – one per traffic source or creative. A Facebook image ad gets its own Ad Group. Your newsletter mention gets another. Your BookBub feature gets another. Each generates a unique tagged link.
Paste that link as the destination URL in your ad or email. Don’t run it through a link shortener like bit.ly. Those strip attribution parameters, and your data disappears.
The Five Metrics That Tell You What’s Actually Happening
The Attribution console gives you a lot of numbers. Most of them are noise. These five tell the real story:
Clicks – how many people clicked your link.
Detail Page Views (DPV) – how many of those people actually loaded your Amazon product page. If this is much lower than your click count, readers are bouncing before the page loads, usually on slow mobile connections.
Purchases – direct sales from that traffic.
KENP Read – pages read by Kindle Unlimited borrowers who came through your link.
New-to-Brand (NTB) – buyers who haven’t purchased from you in 12 months. This is your audience growth number.
The gap between Clicks and DPV is the single most underrated diagnostic. A 30% drop-off there means your ad is bringing people in, but something is killing them on arrival – slow load, wrong store, broken link.
The Meta Gap (Why Facebook Lies to You)
Here’s something Facebook doesn’t want you to think about. When a mobile user clicks your Facebook ad, they’re often redirected into the Amazon app to complete the purchase. Facebook’s tracking pixel cannot follow them there. The sale vanishes from Meta’s reporting.
Amazon Attribution can see it, because it tracks at the Amazon account level, not at the browser level.
The practical result, commonly reported by authors: Amazon Attribution tends to capture 30–40% more sales than what Meta Ads Manager shows for the same campaign. This is important. If you’ve been scaling back ads because Meta says they’re unprofitable, they may actually be working. Run Attribution tags on your Meta campaigns for at least two weeks before drawing conclusions.
Naming Conventions Before You Have Too Many Links
Set up a consistent naming system from day one. Once you’re running five campaigns across three platforms, you will not remember what “campaign4_test2” means.
A clean format looks like this:
US_FB_BOOKNAME_B1_G101_C102_A103
Breaking that down: the country, the platform, the book title (abbreviated), which book in series, the graphic version, the copy version, the audience segment. Every element has a code. The name in your Attribution console should exactly match the name in your ad platform – because then you can use a VLOOKUP to join your ad spend data with your royalty data in a spreadsheet.
That join is where your actual ROI lives.
Kindle Unlimited Authors: Read This Carefully
KENP data is slow. Kindle devices read offline and only sync to Amazon when they reconnect to Wi-Fi. A reader who borrows your book on Monday and reads the whole thing on a train might not have their pages register in your Attribution dashboard until the weekend.
Don’t measure KU campaigns day by day. Give it a full week before drawing conclusions.
Also: the “Estimated KENP Royalties” figure in your dashboard uses last month’s per-page rate as a proxy, because the actual rate isn’t published until the 15th of the following month. Treat those royalty estimates as ballpark figures and reconcile them against your actual KDP report at month-end.
The Algorithmic Side Effect Nobody Talks About Enough
External traffic that converts doesn’t just earn you a sale. It sends a signal to Amazon’s search algorithm that your book is relevant to readers beyond the platform.
Sales velocity improves your Best Seller Rank. BSR improvement leads to better organic placement. Better organic placement means free visibility – a reader finding your book without any ad spend behind it.
There’s also a “halo” effect in the attribution data itself: the dashboard tracks all units sold within the 14-day window, including other books in your catalog. An ad for Book 1 that results in a borrow of Book 3 is a win. You’ll only see it if you’re measuring.
The 2026 View-Through Change (If You Run Display Ads)
In January 2026, Amazon changed how it measures “view-through” conversions – meaning sales that happen after someone saw your ad but didn’t click it. The old model credited any purchase within 14 days of an ad view. The new model uses AI shopping signals to decide if the view actually influenced the purchase.
The result: view-through ROAS dropped 15–30% on paper for Sponsored Brands and Sponsored Display campaigns. This is probably a more accurate number, not evidence that your ads stopped working.
If you run those ad types and your numbers look worse than last year, this is likely why. Amazon added a “Purchases (All Views)” metric that shows the old calculation alongside the new one, so you can compare and adjust your benchmarks accordingly.
Click-based attribution for off-platform links – the type covered by this guide – is unchanged.
The Most Common Mistakes
Running one attribution link for an entire ad campaign instead of one per creative. You’ll never know which image or headline actually converted.
Not waiting long enough for KU data to settle before calling a campaign dead.
Using bit.ly or other shorteners on attribution links.
Trying to track the whole catalog in one campaign. Isolate the book you’re promoting.
Checking daily instead of weekly. Attribution is a 14-day instrument. Daily readings create anxiety, not insight.
Attribution doesn’t make your marketing work. It shows you which parts already do. That’s the whole job – stop guessing, start seeing, and put more money behind what the data actually confirms.
I'd never heard of an Amazon Attribution link. Many thanks for an extremely helpful heads-up, Marcin! 🙏