What I Learned From Breaking My Own Affiliate Clicks While Fixing a Website

The Small-Site Experiment That Broke the Dashboard

I made a website better and my Amazon affiliate clicks fell off a cliff.

That sounds backwards until you remember that a small website is not a vending machine. It is a working system where search traffic, reader intent, affiliate links, AdSense approval, internal linking, content quality, and monetization strategy all pull against each other.

Fix one part and another part can wobble.

Or, in plain English: I adjusted the site, and I broke something.

Not completely. Not fatally. But enough to make the dashboard look ugly and enough to make me question whether I had helped the site or just made it more polite while also making it worse at earning money.

This was not an old authority site with years of traffic.

The site was started in January as an experiment. More on that in a later post, because the site itself belongs in the larger Field Notes trail of what actually happens when small web properties move from idea to traffic to monetization.

The basic idea was to build a practical information site, get it indexed, test search demand, and then layer monetization in a way that matched the reader’s intent. That is the slow side of online income, closer to Money for the Future than “make money today” advice.

The original website monetization plan was simple:

The theory made sense.

AdSense would catch the general readers who came for information but were not ready to buy anything. Amazon Associates would catch the readers who were researching normal consumer products. Direct affiliate referrals would eventually handle the bigger-ticket decisions, where one good lead or sale could be worth far more than a pile of display ad impressions.

That structure fits the larger ABC-eFlow Method: match the money path to the actual stage of the work, not the fantasy version of it.

That was the plan.

The execution got messier.

The plan was AdSense approval first. After repeated denials, the site was changed to look less “Amazon-first.” That meant removing affiliate links from some pages, reducing product blocks, cleaning up diagnostic content, and trying to make the site read more like a useful resource than a buying funnel.

Some of that cleanup was correct.

Some of it probably went too far.

The Amazon click drop was not random. It was at least partly self-inflicted. I had removed links while trying to make the site more acceptable for AdSense. That included some cleanup work after earlier April changes where specific product links had been added to sizing pages.

That is the ugly but useful part.

The early April work showed that specific product links on buyer-intent sizing pages could produce better affiliate behavior. Then the later AdSense cleanup weakened that path.

So yes, I improved parts of the site.

I also cut some of the wires that made the money side work.

The Blunt Lesson

More clicks are not always better. Better clicks are better. That sounds obvious until you are staring at an affiliate dashboard and watching the numbers dry up like a houseplant you swore you were going to water this time.

Here is the pattern I saw in the Amazon Associates data:

The biggest click day made nothing.

The best earnings day came from only five clicks.

That matters.

Traffic is not the same as buyer intent. Affiliate clicks are not the same as revenue. Product links are not magic buttons. A website can generate clicks because people are curious, confused, misdirected, or just poking around.

That may look good in the dashboard. It does not pay.

A smaller number of affiliate clicks from readers who are actually ready to choose something can be worth more than a pile of junk clicks from the wrong page.

Search Visibility Was Not the Problem

The confusing part was that the site was not simply dying.

During the same general period, Google Search Console showed 7.79K impressions, 17 clicks, a 0.2% click-through rate, and an average position of 24.7.

That is not amazing traffic. Nobody is retiring to a beach hut on 17 clicks.

But for a new site started in January, it was movement in the right direction.

That matters because the Amazon drop was not happening because Google stopped showing the site. Search visibility was improving while affiliate clicks were falling.

That points to a different problem. The issue was not only traffic. The issue was the path from search visitor to product action.

I Did Not Just Lose Clicks. I Changed the System.

This is where small website work gets irritating. When you edit a web property, you are rarely changing one clean variable.

You think you are doing one thing:

“I am cleaning up affiliate links.”

But you may actually be doing six things at once:

• changing where the reader exits
• changing which pages carry buyer intent
• changing how Google Search understands the site
• changing how AdSense reviews the site
• changing internal link flow
• changing whether the site feels helpful or commercial

That makes the data harder to read.

• If Amazon clicks fall after a cleanup pass, that does not automatically mean the cleanup was wrong.
• It may mean you removed junk clicks.
• It may mean you removed useful buying paths.
• It may mean both.

In my case, it was probably both.

Some links should have been removed. Diagnostic and educational pages should not feel like product funnels. But some buyer-intent pages had earned their product paths. Removing or weakening those links did not make the site smarter. It made the buying path harder to follow.

That is the part the “passive income website” crowd usually skips. Small web properties do not produce clean lab data. They produce noisy little signals, and you have to interpret those signals without rebuilding the whole site every time one dashboard looks ugly.

Specific Product Links Beat General Affiliate Sludge

One important detail: the better Amazon activity did not happen in a vacuum.

The week before the April 14–15 activity, I added more specific product links to sizing pages instead of relying only on broad marketplace searches.

That matters because sizing pages are closer to buyer intent.

A reader on a sizing page is not just learning basic information. They may already be asking:

“What size unit do I need?”

That is a much better place for a specific product path.

A general Amazon search link can be useful in some spots, but it often punts the reader into the swamp. A specific product link, used carefully, gives the reader a cleaner next step.

Not because every page should sell.

Because the right page should help someone act.

The April 29 result also makes more sense in that context. It came after the site had more specific buying paths on pages where the reader was already closer to choosing.

That does not prove causation with a dataset this small. This is not some grand scientific paper, and I am not wearing a lab coat.

But it is a useful signal.

Specific product links on buyer-intent pages appear to be more valuable than general Amazon links scattered across informational content.

AdSense, Amazon, and Direct Affiliates Do Not Reward the Same Behavior

The original website monetization strategy had three layers.

AdSense was supposed to monetize the reader who wanted information but was not ready to buy.

Amazon was supposed to monetize the reader who had moved from “what is the problem?” to “what do I need?”

Direct or niche affiliate programs were supposed to come later for high-dollar decisions where a serious product match matters more than a generic marketplace link.

That structure still makes sense.

The problem is that the incentives do not line up cleanly.

Amazon rewards product exits.

AdSense wants the site to look like a useful publisher property, not a thin affiliate site dressed up with a few paragraphs of explanation and a trench coat.

Direct affiliates usually want buyer-intent content, but those pages can look commercial if the site does not already have enough trust and informational depth around them.

So the site ended up caught between goals.

Make the pages too commercial, and AdSense may reject the site for low-value content or affiliate-heavy signals.

Make the pages too sterile, and Amazon clicks fall because readers no longer have clear buying paths.

That is not necessarily a broken business model.

It is a sequencing problem.

Page Type Matters More Than Sitewide Rules

One of the biggest mistakes in affiliate website monetization is treating monetization like a sitewide setting.

It is not.

Different pages have different jobs.

This sounds simple, but it gets messy fast.

A page that helps someone diagnose a problem should mostly diagnose the problem. It probably should not push product links before the reader knows what they actually need.

A sizing page is different. That reader may already know the problem and may now need help choosing the right class of product. At that point, a clean Amazon affiliate link or direct affiliate referral can be part of the usefulness of the page.

Same site.

Different reader stage.

Different monetization path.

That distinction matters more than a generic rule like “remove affiliate links” or “add more product links.”

Removing Affiliate Links Is Not Neutral

There is a lazy version of AdSense cleanup that says:

“Take out the affiliate links until AdSense approves.”

That may help in some cases, but it can also break the business model.

The better question is not:

“Should this site have affiliate links?”

The better question is:

“Does this specific link belong on this specific page at this specific point in the reader’s decision?”

If the answer is no, remove it.

If the answer is yes, the link may be part of the page’s usefulness.

A reader who already knows they need a certain product class should not have to start over with a vague marketplace search. At that point, a practical buying path can save time.

But a reader still trying to understand the problem does not need product pressure.

They need clarity.

This is where I overcorrected.

I was trying to make the site less affiliate-heavy for AdSense. That made sense on diagnostic pages. It made sense on pages where the reader was still trying to understand the issue.

But when that same cleanup weakened specific product links on sizing pages, it started to hurt the part of the site that was supposed to help buyers.

That was the mistake.

The March 29 Problem

The 33-click day with zero orders is useful because it looks good at first and then falls apart.

Thirty-three Amazon clicks sounds like progress on a small site.

It was not.

It was probably noise.

Possible explanations:

• the link placement attracted curiosity clicks
• the page was not buyer-intent enough
• the product path did not match the reader’s problem
• the link appeared too early in the page
• readers clicked because they were confused, not ready
• the product did not fit the promise of the content

Whatever the exact cause, the result was the same.

Thirty-three clicks.

Zero orders.

That is not a monetization win.

That is a warning light.

The April 14 and April 29 Lessons

April 14 and April 15 were different.

There were only five clicks across those two days, but one item was ordered and then shipped. That produced $5.79 in earnings.

April 29 was better.

Five clicks.

Two ordered items.

Actual affiliate earnings.

That is the kind of affiliate behavior a small website needs.

Not button spam. Not “top 10 best” filler. Not fake urgency. Not a thin affiliate page with a buying table glued to the top and a few paragraphs of generic explanation underneath.

Just a reader with a real problem reaching a page that gives them a reasonable next step.

That is the target.

The goal is not more clicks from every visitor.

The goal is better clicks from the right reader at the right point in the page.

And in this case, the right point appeared to be the sizing pages where specific product links matched the page intent.

The Trap: Fixing Everything at Once

The worst thing you can do after a weird data swing is panic-edit the entire site.

I know this because that is always the temptation.

Amazon clicks fall?

Add links back everywhere.

AdSense rejects the site?

Strip all monetization.

Google impressions rise but clicks stay low?

Rewrite every title.

One page ranks?

Turn it into a giant hub.

That is how you turn a small site into rubble.

The smarter move is slower and more annoying:

1. Identify the page type.
2. Decide the job of that page.
3. Match monetization to that job.
4. Change fewer things at once.
5. Watch the next signal without pretending one week proves everything.

That is not exciting.

It is also how real website tuning works.

A small website is not passive income in the way people usually sell it. It is one version of a side-income system, but it sits much closer to the long-game category than the quick-cash category described in How Side Gigs Generate Income.

The Second Lesson: AdSense May Not Deserve First Priority

Here is the part that changed my thinking.

After making the site more AdSense-safe and more Amazon-poor, AdSense rejected it again anyway.

That is a useful slap in the face.

The original plan put AdSense first because it felt like the clean foundation. Get approved, let display ads catch the general readers, then layer Amazon and direct affiliates more carefully on buyer-intent pages.

That still sounds logical.

But the math is not as friendly as the theory.

An Amazon sale on a decent home product can beat a pile of display ad impressions. A direct affiliate referral on a higher-dollar product can be worth $50 to $100 or more. It takes a lot of ad views to equal one serious buyer action.

So the question changed.

It stopped being:

“How do I make the site safe enough for AdSense?”

And became:

“Why am I weakening better monetization paths to chase the weakest one?”

That does not mean display ads are useless.

They can still catch casual readers who will never buy anything. They can still add background revenue once traffic exists. They may still belong on the site eventually.

But they no longer deserve to drive the whole strategy.

Especially not if the cost is making useful buyer pages worse.

Nobody likes landing on a recipe page and fighting through six ads, two popups, a sticky video, and a personal essay just to find out how long to boil water.

I am not anti-story. I am not even anti-ad.

But when I search “how long to boil water,” I do not need a full spiritual memoir about someone’s grandmother, a sticky video ad, a newsletter popup, and three autoplay blocks before I find out that water boils when it boils.

That is not the experience I want to build.

If AdSense approval requires turning the site into something less useful or stripping out legitimate buying paths, then AdSense needs to move down the priority list.

The Revised Monetization Priority

The better order now looks like this:

That is a major shift.

AdSense was supposed to be the first monetization gate.

Now it becomes optional.

Not because display ads are evil. Because the economics are weaker, and the approval process can push the site in the wrong direction.

My Current Rule for Website Monetization

For a small content site, I would rather have fewer affiliate links in better places than more links scattered across the site.

The working rule is:

Diagnose first. Decide second. Sell last.

That means:

• diagnostic pages stay clean
• explanation pages stay useful
• decision pages may have limited product paths
• sizing and selection pages carry the main buyer links
• direct affiliate paths belong where higher-dollar product decisions are real
• product links should be grouped, not scattered
• no fake urgency
• no pretending every reader is ready to buy

That structure may reduce total affiliate clicks.

That is acceptable if the remaining clicks are more qualified.

It also means I need to stop treating every monetization cleanup as an AdSense appeasement ritual.

There is a difference between making a site trustworthy and making it toothless.

I managed to blur that line.

That was the mistake.

What This Means for Side-Income Websites

A small website is not passive income in the way people usually sell it.

It is not:

“Write article, add link, collect money.”

It is closer to:

“Build a system, watch the signals, break part of it, fix the wrong part, learn something, then make a cleaner change next time.”

That is less exciting, but it is more honest.

The money is not only in publishing pages. The money is in tuning the path between reader intent and the right next action.

Sometimes that next action is an affiliate link.

Sometimes it is an internal link.

Sometimes it is no link at all because the reader is not ready.

That judgment matters more than the number of buttons on the page.

The Takeaway

The highest-clicking Amazon day did not make money.

The best revenue day came from a handful of better clicks.

Search visibility was improving, so the Amazon drop was not just a traffic problem.

The early April product-link work suggested that specific product links on buyer-intent sizing pages could work.

Then, while trying to make the site cleaner for AdSense, I removed or weakened some of the paths that helped buyers act.

And after all that, AdSense rejected the site again anyway.

Lovely.

That changed the lesson.

The goal is not to make every monetization network happy.

The goal is to build a useful site with a sane money path.

For this property, that probably means:

• clean informational pages
• strong search visibility
• no affiliate links on diagnostic content
• buyer links where the reader is actually choosing something
• specific product links where they genuinely fit
• direct affiliate paths for higher-dollar decisions
• AdSense only if it fits without distorting the site

This is why I keep separating quick money, stable extra income, and long-term web property work. They are not the same game. A small website can become useful, but it usually belongs in the future-income bucket, not the emergency-cash bucket.

I made a real improvement.

I also overcorrected.

Both things can be true, which is deeply annoying and usually how learning works.

A small website is a working system.

Tune it like one.

But do not let the lowest-value monetization channel dictate the whole design.