Your Amazon affiliate site is not making money. Find out which link in the chain is broken.
“Not making money” has five completely different causes with five different fixes, and the advice you get depends entirely on who you ask. Work down the funnel instead, find the stage where your numbers collapse, and fix that one.
The short version
- Find the stage that is failing before changing anything. Traffic, click rate, conversion, and earnings per sale fail for entirely different reasons.
- If you have under about 1,000 monthly visitors, nothing is broken. You do not have enough traffic for the maths to work yet, and no amount of optimisation will change that.
- If earnings fell in 2026 while traffic held, look at the April attribution change and at dead products before you blame your content.
- If traffic fell in 2026, look at AI Overviews. Content that a summary box can answer lost most of its clicks.
- The most common single cause is the least dramatic: too few pages, targeting terms that were never going to rank.
The unhelpful thing about this question is that every answer you find is somebody’s favourite fix. Write more content. Improve your product boxes. Change niche. Get better at SEO. Each is right for one specific failure and useless for the other four.
So do not start with a fix. Start by finding out where your numbers stop looking like a working site.
The funnel, with numbers to compare yours against
Every affiliate site is the same four multiplications. Here is a rough working model for a healthy review site — not a promise, and not benchmark data, but close enough to show you which stage is out of line.
How to get your four numbers. Visitors from Search Console or analytics. Affiliate clicks from Amazon’s reports, or better from your own click tracking, which is not blocked and not delayed. Orders and earnings from Amazon’s earnings report. Divide each stage by the one above it.
Stage 1 — not enough traffic
This is the most common answer by a wide margin, and the least welcome, because the fix is slow.
Under roughly 1,000 monthly visitors, your site is not broken — it is early. At the model above, 1,000 visitors is around $70 a month. There is no product-box tweak, no plugin and no prompt that turns 1,000 visitors into a living. Optimising conversion at this stage is rearranging a shop that has no customers.
The usual causes, in order of how often they turn out to be the real one:
- Too few articles. Twenty posts is a start, not a site. Most sites that earn have well over fifty pages, and most of those pages earn nothing — the winners are a minority you cannot pick in advance.
- Keywords that were never winnable. “Best laptop” will not rank for a new site. “Best laptop for architecture students under $1,200” might.
- Not enough time. New sites commonly take six to twelve months to rank for anything meaningful. Checking weekly and despairing is normal and unhelpful.
- Content an AI Overview answers. New in 2026 and it matters — pages answering questions that a summary box handles in three sentences lost much of their click-through. Content built on hands-on testing held up far better. This changes which topics are worth writing at all, and it is covered in the 2026 niche guide.
What to do: keep publishing, target longer and more specific terms, and give it months rather than weeks. If you have fewer than 30 articles, that is the whole answer and the rest of this page can wait.
Stage 2 — traffic, but nobody clicks through
You have visitors and almost no affiliate clicks. Two very different explanations.
Your traffic is informational, not commercial
Someone searching “how to season a cast iron pan” wants an answer, not a purchase. Someone searching “best cast iron skillet for induction” is shopping. Both are useful — but only one converts, and a site made entirely of the first will always look broken.
Check the queries in Search Console. If your top pages are all how-to questions, your content is fine and your keyword mix is wrong. The fix is to add commercial pages, not to bolt product boxes onto guides where they do not belong.
Your links are hard to find or easy to ignore
- No recommendation above the fold, so the reader has to scroll to learn what you actually suggest.
- Plain text links where a button belongs.
- A ten-product roundup that gives the reader a research project instead of an answer. Three well-argued options beat ten listed ones.
- No clear verdict. Readers want to be told which one to buy.
Stage 3 — clicks, but no orders
This is where the silent, mechanical failures live, and where most people never think to look.
Dead products
Discontinued or out-of-stock items return a healthy 200, so no link checker flags them. The reader lands, cannot buy, leaves. The biggest invisible leak on any site over a year old. → how to find them
Untagged links
A missing tag= means the sale happens and you are not paid. Nothing looks wrong. → how to check
International readers
A UK reader on a .com link mostly leaves, and if they buy on their own store you earn nothing. → how to fix it
The 24-hour window
Amazon’s cookie lasts a day. Expensive, considered purchases lose a real share of conversions to deliberation, and there is nothing to fix — only to expect.
Work through those four in that order. The first three are all fixable in an afternoon and all of them are invisible until you look deliberately.
Stage 4 — orders, but the money is tiny
People are buying and the total is still depressing. This is an arithmetic problem, not a marketing one.
- You are in a 1% category. Grocery, health and personal care pay 1% — sixty cents on a $60 order. No amount of traffic rescues that.
- Your products are too cheap. 3% of $30 is ninety cents. 3% of $400 is twelve dollars. Same article, same effort, thirteen times the return.
- You lost the halo income. Until April 2026 you earned on whatever the shopper bought. Now it is the product you linked and its variants. If your earnings dropped while orders held steady, this is very likely why. → the 2026 rate and attribution guide
The highest-leverage change available to most sites. Move up the price ladder. Writing about $300 products instead of $30 ones takes the same effort, ranks about as easily, and pays ten times more. The reason most people do not is that cheap products feel easier to recommend — but the reader researching a $300 purchase is exactly the reader who reads a whole article first.
Special case — it used to work and now it does not
A different problem from never having earned, and 2026 gave it several new causes.
| Symptom | Likely cause |
|---|---|
| Traffic fell, earnings followed | An algorithm update, or AI Overviews taking clicks from answerable queries. Check Search Console impressions versus clicks — flat impressions with falling clicks is the AI Overview signature. |
| Traffic flat, earnings fell | The April 2026 attribution change, or dead products accumulating. Check dead products first — it is faster and more often fixable. |
| Clicks flat, orders fell | Dead or unbuyable products on your best pages. |
| Everything fell at once | Something structural: links stripped by a plugin update, a caching change serving old links, or account trouble. → account status |
| Product boxes are blank | API access withdrawn. Eligibility is a rolling requirement now. → API access denied |
What to do first
In order. Stop when you find something.
- Work out your four funnel numbers. Thirty minutes, and it tells you which of the sections above applies to you.
- If traffic is under 1,000 a month: publish more, target longer keywords, wait. Nothing else matters yet.
- Audit your top 20 pages for dead products. An hour, and on an older site it usually pays for itself immediately.
- Check every Amazon link carries your tag. Untagged links are pure loss.
- Check your international traffic share. Over 25%, fix it this week.
- Look at your average order value. If it is under $50, move up the price ladder.
- Then, and only then, think about redesigning product boxes.
The uncomfortable one. Sometimes the honest answer is that the site is too young and there is nothing to fix. Most affiliate sites that fail do so because they were abandoned in month four, when the work was done and the results had not arrived yet. If your numbers say “keep publishing”, that is a real diagnosis, not a consolation.
Part of The Amazon Affiliate Site Playbook — the whole job in six stages, from choosing a niche to keeping a mature site earning.
Find the leaks without checking 300 posts by hand
Ama Affiliate Pro finds products that have stopped being buyable, links missing your tag, and pages earning nothing despite traffic — plus first-party click tracking so you can see your real funnel instead of Amazon’s delayed version.
Start a 14-day trialCommon questions
How long before an Amazon affiliate site makes money?
Typically six to twelve months before meaningful search traffic arrives, and longer before it is a living. First commissions often come in months three to six with 20–40 published articles. If you are three months in with fifteen posts and no earnings, that is the expected result rather than a fault.
How much traffic do I need to make $1,000 a month?
On the rough model on this page — 10% of visitors clicking, 7% of clicks converting, $10 average per order — around 15,000 monthly visitors. Higher-priced products cut that figure substantially; a site averaging $25 per order needs roughly 6,000. This is why average order value matters more than almost anything else.
Why do I get clicks but no sales?
Check three things in order: whether the products are still buyable, whether your links carry your tag, and where your readers are in the world. After that, remember the cookie lasts 24 hours, and that since April 2026 you only earn on the product you linked and its variants rather than on the whole basket.
My earnings dropped in 2026 but my traffic did not. Why?
Most likely the April 2026 attribution change, which stopped paying commission on unrelated products bought after the click. That income was invisible in reports, so its loss looks inexplicable. The other candidate is dead products accumulating on your best pages — check that first, because it is faster to verify and usually fixable.
Should I change niche if my site is not earning?
Rarely, and not before checking the funnel. Changing niche restarts a clock that takes six to twelve months, and if the real problem was too few articles or too-cheap products you will reproduce it in the new niche. Change if your category pays 1%, or if the products are consistently under $30. Otherwise fix the stage that is failing.
Is Amazon affiliate marketing still worth it in 2026?
For sites built on genuine first-hand experience with mid-to-high priced physical products, yes — Amazon converts better than almost any other programme and readers already trust the checkout. For thin roundups assembled from other people’s reviews, much less so; that model lost ground to both AI Overviews and the attribution change. The rates are low and always were; the conversion rate is what makes it work.
How many articles do I need?
Think in terms of covering a topic rather than hitting a number, but as a practical floor: 30 before you can judge anything, 50 to 80 before a narrow niche is genuinely covered. Expect most of your revenue to come from a small minority of pages you could not have identified in advance — which is the actual argument for volume.
Should I add display ads as well?
Once you have consistent traffic, yes — ads pay on visitors who never click an affiliate link, which is most of them. Be careful on high-intent pages, where ads can pull readers away from a click worth far more than the impression. A common approach is ads on informational content and none on the commercial pages.