Amazon took away your
affiliate data. Here's what
you can still track.

If you run a content site with Amazon affiliate links, 2026 has not been kind to your reporting dashboard. Two changes — one in March, one in April — closed off visibility that many creators had quietly depended on for years. The reaction in creator communities ranged from frustrated to resigned.

This post is a practical breakdown of what happened, what it actually means for your data, and how a redirect-based tracking approach can recover some of what's now missing. We'll be specific about what it can and can't do, because the honest answer matters more than an oversell.

What Amazon changed, and when.

The changes came in two waves:

The combination is what makes it acute. The SubID gap always existed, but creators tolerated it because Amazon's own dashboard at least showed order-level activity. Now that order-level view is gone too, leaving many affiliates with nothing but a topline earnings number and a click count that doesn't tell them which article or which link drove it.

What's gone forever vs. what's still recoverable.

It's important to separate these two categories. Some things are sealed off at the platform level. No third-party tool can reconstruct them, and anyone claiming otherwise is either confused or misleading you.

The "still recoverable" column lives entirely on your side of the click — in the moment between a reader tapping a link and arriving at Amazon. That's the window where a redirect-based tracker operates. The "gone permanently" column lives inside Amazon's walls, and no external tool has access to it.

Understanding this distinction matters because it shapes what you should realistically expect from any tracking solution, including this one.

How redirect-based tracking works — and why it matters.

The core idea is simple: instead of linking directly to Amazon, you link to a short URL you control. When a reader clicks that link, a server intercepts the request, logs everything useful about it, and immediately redirects them to Amazon. The reader notices nothing. You get the data.

The "probably" in that last step is important. Without a conversion postback from Amazon, attribution is always probabilistic, not definitive. An article that generates 400 clicks to a $50 product in a week where you earned $60 from that product is likely responsible — but you can't be certain. Honest tools communicate this. Dishonest ones don't.

The ToS question you're probably already thinking about.

Amazon's ToS explicitly prohibits "cloaking" affiliate links — presenting a URL that hides the fact that the destination is Amazon. This is why tools like ThirstyAffiliates operate in a gray area: a link like yoursite.com/go/vacuum doesn't tell the user they're going to Amazon before they click.

The redirect approach that keeps you compliant is to include the destination domain in the URL path itself. A link like go.linkreflo.com/amazon-com/dyson-v15 makes it clear, before the click, where the reader is headed. The tracking happens at redirect time, not by obscuring the destination.

This is not legal advice and you should read Amazon's current ToS yourself. But it's the design principle that informs a compliant redirect tracker, and it's meaningfully different from link cloaking.

What this looks like in practice.

We've been running this approach on A Pom's World — a pet content site — since April 2026. We replaced every raw Amazon link in our articles with tracker links and started accumulating click data. A few things we've learned in the first weeks:

Bot traffic is substantial. A significant portion of early "clicks" are crawlers, not readers. A properly built tracker classifies these at capture time so your human click counts are accurate and your export is pre-filtered.

Source attribution is genuinely useful. Knowing that 70% of clicks on a specific product link come from one article — not distributed across ten — changes how you think about internal linking and content prioritization. That signal was there before, hidden. Now it's visible.

Data accumulates slowly at first. Two to three weeks of real click data is the minimum before patterns emerge. The first week mostly tells you that the system works. Month two is when it gets interesting.

Should you do this now, or wait for a better tool?

The honest answer is: start now with whatever you can, even imperfectly. Click data is time-stamped. Data you don't capture today is gone. If you wait three months for the perfect attribution dashboard and then set up tracking, you'll have three months less signal when you try to analyze your articles' relative performance.

The minimum viable version is: replace your highest-traffic article's Amazon links with tracked redirects. Pick your five best-performing pieces and do those first. Let four weeks of data accumulate. Then look at which links in which articles are generating actual clicks — and use that to prioritize where you spend your writing time next.

That's not a complicated workflow. It doesn't require understanding Cloudflare Workers or running terminal commands. It requires making the decision to start, and then actually starting.

The reporting changes Amazon made in early 2026 were a setback, not a death sentence. The data that matters most for making content decisions — which articles drive clicks, which links perform, which content is worth doubling down on — was never inside Amazon's dashboard to begin with. It was always yours to capture. You just needed a way to capture it.