Making money with this

The truth first

What follows is a calculation framework, not a promise. Figures vary enormously by niche, reader country and season. Treat every range quoted here as an order of magnitude to verify with the platforms themselves.

Making money in 2026, in the AI era

This has to come before anything else, because it changes the plan for the whole page: Google increasingly answers instead of the websites it used to send you to. The reader gets the answer on the results page and never clicks. That is not a feeling, it is measured, and the 2026 numbers are harsh.

What still works — and why

The good news: the traffic loss is not spread evenly. It hits one type of page very hard and barely touches the others. Here is the measured split.

Sources: First Page Sage 2026, Contently 2026.

The five levers, plainly

1. Write for purchase intent, not for curiosity

   “Best X for Y”, comparisons, reviews, alternatives, “X vs Y”: those pages keep their traffic because buying requires a click. A “what is X” page is replaced by the AI. Same writing effort, completely different value.

2. Pick a local or service niche

   Local services lost 6.8% while publishers on average lost 38%. An AI cannot quote you a plumber's price in your city, or their availability. It is the most protected ground there is right now — and the least used by autoblogs.

3. Publish what a model does not have: your own data

   A price you checked yourself, a test you ran, an up-to-date rate table, a dated first-hand report. A model cannot invent a figure it has never read: that is the one thing it has to come and fetch from you, and cite.

4. Get cited by the AIs, not just ranked by Google

   A brand cited in an AI Overview earns 35% more organic clicks on the same query than one that is not. And 75–81% of ChatGPT and Claude citations point at third-party “best X” pages — precisely the kind of page you can write.

5. Add a tool, not just text

   A calculator, a comparator, a simulator, an interactive checklist: under 3% of those searches trigger an AI Overview, because a text box cannot do the sum for the reader. It is also what brings people back.

Getting paid by AI for your content: where it actually stands

You have probably seen Cloudflare's announcement: charging the AI crawlers that come and read your site. The idea is real, the plumbing exists, and it is simple to understand.

How it works. Cloudflare sits in front of your site. When an AI crawler asks for a page you have three buttons: allow, block, or charge. Choose charge and the crawler gets a 402 Payment Required response — an HTTP code invented thirty years ago for exactly this and almost never used — carrying your price. If it agrees, it re-sends the request cryptographically signed, gets the page, and Cloudflare collects the money and pays you. You set one flat price for the whole site, per request.

Our position, straight: this is an option to keep open, not a revenue stream. Putting your site behind Cloudflare is free, takes half an hour, protects you along the way, and puts you in the queue if the day comes when AI companies really pay. But do not change a single decision in your plan because of it, and do not block AI crawlers while you have nothing to sell: in 2026, being cited is measurably worth more (+35% clicks) than the toll is.

Sources: Cloudflare blog (mechanics and private beta), TechCrunch 2026, Search Engine Land (RSL).

The two models

In practice, the sites that work do both: ads monetise informational traffic, affiliate monetises buying traffic. You do not have to choose permanently.

Affiliate: the maths, explained very simply

You talk about a product. Someone clicks, buys, and the seller pays you a percentage. That percentage is the commission, and it changes everything.

The formula:

revenue = visitors × outbound click rate × conversion rate × order value × commission

Real commission rates, by product type

This is the most important table on the page. Look at the last column: same traffic, same work, twentyfold difference.

The same site, two different niches

3,000 visitors a month in both cases. Only the niche changes:

Nobody worked harder in the right-hand column. The decision was made before the first article was written — which is exactly what the niche page covers.

The levers, most powerful first:

1. The commission rate. From 4% to 50%, everything multiplies by twelve without one extra visitor. The strongest lever, and it is decided by the niche.
2. Recurring revenue. A subscription that pays you monthly beats a one-off sale, always.
3. Order value. An $800 product is worth ten $80 products.
4. Purchase intent in the articles. A comparison converts ten to twenty times better than a “what is X” article.
5. Traffic. The slowest — and the one everybody focuses on first.

Ads: the thresholds

Ad networks sort by entry threshold and payout quality. Thresholds change — always check current requirements on the network's own site:

Ad revenue is measured as RPM: revenue per 1,000 pageviews. It varies by a factor of ten depending on niche (finance and insurance pay far above hobbies), reader country (North America and Northern Europe pay several times what the rest of the world does) and season (November–December is a sharp peak, January a trough).

Practical consequence: do not build your plan on an RPM you found in a blog post. Put AdSense up, measure your real RPM over a full month, and redo the calculation with your own number.

Three scenarios: bad, good, excellent

Here is what one site, after 12 months looks like, publishing 30 quality articles a month. The three cases are not three levels of effort: it is almost always the niche choice that decides which one you land in — and since 2026 the first test of that choice is whether AI already answers instead of your articles.

Two things jump out, and they are the only two to remember:

1. The “digital product” row crushes the “ads” row everywhere. Same site, same traffic, same work — only the niche and the programme differ.
2. Ads only become interesting on the right of the table. Below 10,000 visitors they will not pay for your coffee: affiliate is what carries a young site.

What moves you from column 1 to column 3: the niche choice (see the full method), real review, purchase intent in the topics, and the patience to let it run 12 months. Nothing else.

Which to choose

Traffic maths

Work backwards from your goal. To aim at $500 a month:

• On ads, at a $10 RPM (an average assumption to be verified): you need 50,000 monthly pageviews. At 3 pages per visit that is ~17,000 visitors.
• On affiliate, in a decently paying niche: 3,000–8,000 well-targeted visitors is often enough.

The gap between those two lines is why a young site almost always starts with affiliate.

How many articles for that traffic? The honest answer has no fixed number: a hundred articles that do not rank are worth zero, and ten well-positioned ones can carry a whole site. Article count is not the variable — the number of articles that rank is.

The real timeline

Costly mistakes

1. Picking the niche on commission rather than on your own competence. The best-paying niches are YMYL and judged hardest: if you cannot fact-check, you cannot go there — see the risks.
2. Publishing 200 articles before knowing whether the niche works. Publish 20, wait three months, look at the data. You will save the entire niche if it is wrong.
3. Ignoring buying intent. A hundred “what is X” articles pay less than ten comparisons.
4. Not disclosing affiliate links. A legal obligation and a search engine one — both.
5. Putting ads up too early. Below 10,000 visitors they earn nothing and degrade the experience, which degrades rankings.
6. Never updating. A maintained article beats a new one. It is the cheapest and least-used lever there is.

A 12-month plan

1. Month 1 — decide

   A niche you can fact-check. Check that an affiliate programme exists and at what rate before writing anything. This is the cheapest moment to change your mind.

2. Month 1 — 30 articles

   Generated then seriously reviewed, one a day. Including 5 buying-intent pillars. Publish over 3–4 weeks, not in one day.

3. Months 2–3 — measure

   Search Console: what is indexed, what gets impressions? If nothing is indexed after two months, the problem is the niche or the quality, and publishing more will not change it.

4. Months 3–6 — extend what works

   Double down on topics getting impressions. Delete the rest. Put AdSense up to measure your real RPM.

5. Months 6–12 — go deeper

   Update the articles that perform: more profitable than writing new ones. Add affiliate where buying intent exists.

6. Month 12 — decide again

   Does the site earn enough to justify the time? If yes, the 10-site model becomes relevant. If not, the niche was wrong — and you learned it for tens of dollars instead of thousands.

The cost of 10 sites → Best practices