Making Money from a Blog Now That Clicks Are Running Out

It used to be simple. The more you wrote, the more visible you were, the more readers showed up, the more you earned. For years, blog monetization ran on this accumulation equation built on ads and clicks, and sites popped up that talked about literally anything: everyone had something to say, and everyone could find a way to get paid for saying it.

Then the web started to change. Social platforms took a slice of attention, Google started keeping searches on its own page, and production costs went up. You could still make it work, as long as you ran faster.

Now we’ve hit the breaking point. In the Italian market, nearly two out of three searches end without a single click, AI Overviews satisfy the searcher’s need on their own, and generative systems have become direct competitors to the people who write that content in the first place. The old model isn’t dead, it simply no longer covers the costs for almost anyone. Alternatives are needed, and some have already found them, in Italy too.

Monetizing a blog has never been just one thing

Saying “my blog makes money” has always meant economically different things that, for a long time, stayed indistinguishable from one another.

• Advertising sells the ad space every open page makes available: money comes in the moment someone reads, and it scales with how many pages get opened.
• Affiliate marketing pays a commission when someone clicks through to a store and buys: the payout lands days later, and only if the purchase gets attributed to you.
• Sponsored content sells a company your readers’ attention, along with the reputation your domain has built: here the money arrives before publication, and the reader isn’t the customer.
• Company blogs don’t make money on the page itself; they exist to sell software, a product, or consulting services: the revenue shows up downstream, weeks later, and often without any way to trace it back to the article that set it up.

Different customers, different timelines, different costs. What held these different economics together was that they all started from the same point, someone landing on the site, and as long as that number kept growing, improving it lifted all of them.

Distribution has moved to where you don’t collect money

The problem is that traffic has stopped growing, so the business models built on top of it have collapsed as a result. Not because your audience lost interest in what you write: they keep searching for it and finding it, just in different places than your site.

Google displays your information right on its own page, inside boxes and generative summaries, and searchers get what they wanted without moving a finger: in the Italian market, 63.4 percent of browser searches end with zero clicks. A conversational assistant takes the same content and uses it to build an answer your site will never see, and generative systems send back less than 1 percent of total traffic to sites. Some questions, meanwhile, have stopped going through search at all: people look for reviews on TikTok, experiences on Reddit, tutorials on YouTube, where whoever answers has no reason to send anyone your way.

Your work pays off, in other words. The moment when traffic turned into money has come unstuck, because it used to coincide with opening the site, and now that step gets skipped.

That also explains the clash that followed, with European and American publishing groups filing lawsuits against Google and against the companies building generative models, arguing that their articles are being used to build answers that cut traffic to the source.

It’s not the cause of the split, it’s the reaction: however it turns out will matter little to your blog in the short term, while the relationship between content creators and distributors has already stopped being an implicit deal. What matters, when deciding what to do, is that the hit doesn’t land the same way for everyone. In the second quarter of 2026, publishers’ ad inventory dropped by as much as 40%, according to Ozone benchmarks for the US and UK, while an ecommerce site losing the same traffic can see revenue hold up. The difference comes down to which piece of the old mechanism you actually relied on.

Advertising runs on page views, and those are shrinking

Display advertising pays very little per unit, and that single fact explains almost everything Italian online publishing has done over the past fifteen years: sites covering any topic under the sun, twenty-page photo galleries, headlines built for SERP clicks, “read also” boxes halfway through an article. These weren’t bad editorial choices, they were the only lever available when every page view is worth a few cents.

You can do the math with your own numbers, and you need to before anything else: divide monthly costs by Page RPM, multiply by one thousand, and you get the pageviews you need to break even. Page RPM is what you earn per thousand pages loaded, you’ll find it in your reports, and it varies wildly between two sites that look identical. A personal finance blog earns more than an entertainment site with the same traffic, because companies in that space compete for the same audience and bid up the auction price. With 3,000 euros in costs and an RPM of 3 euros, you need a million monthly pageviews. If RPM rises to 6, you only need 500,000; if it drops to 1.5, you need 2 million.

That number tells you how much of your work depends on volume you don’t control. And when it starts to drop, the temptation to add more ad formats shows up right on cue, to squeeze more out of every session: it works until the bill comes due, because the page gets slow, cluttered, tiring, and what you’re eroding is exactly the reason people would come back, which is what should have driven next month’s visits. Salvatore Aranzulla has said he lost a quarter of his traffic. His site runs on advertising and affiliate income, both of which only exist if Google sends people his way: when that flow thins out, everything falls at once.

Within this model, the margin disappears along with the visit, because the visit was the product. People who’ve made a living this way for years are now trying to move the payoff one step further down the funnel, onto the purchase instead of the read.

Affiliate marketing pays you for the sale, not the read

No negotiating, no client to manage, you sign up for a program and get started, earning money from a purchase completed on someone else’s site. That’s the convenience of affiliate marketing, which has made it the most common add-on income alongside banner ads. 

The mechanism is simple: you put a link containing your own ID code in your content, and when someone clicks it and buys on the seller’s site, you get paid, a percentage of the order, a flat fee, or an amount per qualified lead, depending on the program.

Your net earnings track the seller’s margins: software, courses, and services leave more room to pay for referrals than low-margin physical products do. Then there’s the attribution window, which decides whether that work gets paid at all: the link between your article and the purchase lasts anywhere from a few hours to a few months depending on the program, and outside that window you get nothing. For percentage-based programs, the math is outbound clicks times the seller’s conversion rate times average order value times commission, minus returns and cancellations. It looks like a fix for the previous problem, and it’s only half of one: that outbound click still comes from a visit, which is exactly what you’re losing.

When an assistant does the work instead

A comparison page used to have a very simple value: it pulled together, in one place, prices, specs, and reviews scattered across ten different product pages.

Someone shopping for a refrigerator would open your article because it saved them two hours. Now they get that time back elsewhere: inside ChatGPT, people explore and compare products by pulling from merchant catalogs, and the table shows up already filled in. Built-in purchase features currently have limited availability by region, but the comparison function doesn’t.

Affiliate pages that just replicate the seller’s materials without any substantial original contribution, for their part, can fall into what Google calls thin affiliation and end up among the spam policies violations. But the risk has changed in nature: it used to be a penalty, now it’s that the reason to visit them is gone altogether. What still holds up is the work someone actually bothered to do: the cost that shows up after purchase, the incompatibility you discover once you’re using the product, the maintenance that turns out to be a hassle, the trade-off you only notice months later. A review that holds up also tells you who that product is the wrong choice for, and whoever signs it takes responsibility for it.

The same goes for a comparison site that updates prices daily, a database nobody else has, or a curated selection that a niche audience trusts as reliable: anything that requires ongoing work instead of a one-time collection effort. Advertising and affiliate marketing both rely on people who pass through once. When that flow shrinks, the alternative is to stop counting on strangers and start working on the people who come back.

From the occasional visitor to a list you actually own

Working on the people who come back means being able to reach them, and that’s what makes the paid newsletter the most common path these days: you publish part of your content via email, reserve part of it for monthly or annual subscribers, and distribution stops depending on an algorithm.

The real advantage is more specific than it’s usually described. Analytics reports have a category called “direct” that collects sessions without an identifiable source – someone who opened a link from an app, a message, a document – and it tells you how the tool classified that visit, not whether you have a way to reach that person. An email list, on the other hand, contains addresses you can use whenever you decide, within the consent you’ve collected, and that you take with you if you switch platforms: that’s the difference between hoping someone comes back and being able to go back to them yourself. Substack, Ghost, beehiiv, and Telegram all do this job, and they differ in how much they keep and how much they leave in your hands.

• Substack keeps 10 percent of subscription revenue plus payment processing fees, drives subscribers through its network of cross-recommendations between publications, and lets you export your list.
• Ghost and beehiiv charge a flat fee instead of taking a cut of revenue, so which one makes sense for you depends on your list size and the features you need.
• Telegram lets you run paid subscriptions on closed channels and works well where immediacy and a sense of community matter, but there’s a tradeoff that counts: you reach the same audience repeatedly, yet both identity and distribution stay inside the app, and you can’t take your list with you.

Having contact details, though, is still a different thing from getting paid. In the first quarter of 2026, BuzzFeed had more direct traffic on its main site than from Google and Facebook combined, the condition often described as the goal, and it closed that same quarter with $31.6 million in revenue, down 12.4%, and a net loss of $15.1 million – advertising down 19.8%, commerce down 32%, and the only growing line being commissioned content and licensing, up 69.1%.

You still have to do the math with two numbers of your own: how many free subscribers you have and how many of them would pay. With five thousand subscribers, a hypothetical 3% conversion at ten euros a month produces eighteen thousand euros gross a year, before fees, taxes, and churn. That 3% isn’t a benchmark of health, and it changes with price, frequency, how targeted the list is, and how long those people have been reading you for free.

In Italy, the market weighs more than elsewhere: 8% of respondents to the Digital News Report 2026 say they pay for online news, with trust in news down to 32%, four points lower than the previous year. That’s about journalistic news and says little about courses, professional analysis, or niche communities, but if you write for a general audience, that number tells you what you’re up against. The real question is why someone should pay you at all.

Anything Goes, and the Capital That Was Already There

Selvaggia Lucarelli was a recognizable byline long before she launched a newsletter. Twenty years across newspapers, television, books, and social media built an audience that knows what to expect from her and follows her even when they disagree. Vale Tutto channels that relationship into a format that can be monetized directly: subscribers get at least four articles or investigations a month, original documentaries, early access to video podcasts, travel itineraries, and direct access via email.

Substack, in all this, removes the technical friction – publishing, sending, collecting payments – without creating an ounce of willingness to pay on its own. So someone who arrives with an established byline brings trust and attention built elsewhere into the newsletter and can convert from day one, while someone starting without an audience has to build both inside the free tier, on a timeline that depends on how recognizable the byline is, how urgent the need is, and how many other people already meet it. Copying the model without the capital behind it leads where you’d expect: a technically flawless newsletter with almost no paying subscribers.

Readers can pay to keep a piece of work alive that stays free for everyone else

A byline like Lucarelli’s belongs to very few people. But there’s also a model that works in exactly the opposite way: Il Post publishes nearly all its articles unsigned, as an editorial choice, because what comes out is the result of collective work and should be credited to the publication, not to a single author. And it keeps almost all of its output open: two conditions that should make a subscription impossible to sell, since there’s no name to follow and no content to unlock.

Instead, in 2025 the outlet topped 11 million euros in revenue, up 18%, subscriptions grew 22%, and they remain the largest share of revenue.

Paying readers get exclusive podcasts and newsletters, the app, and access to events. But above all, they’re paying for the chance that the newsroom keeps working, in what we call a participatory model.

The Digital News Report captures this too: among Italians who say they’ve paid for online news, 38% cite at least one social or values-based reason – supporting journalism, identifying with that outlet’s values, or letting content stay accessible to people who don’t pay.

L’Ultimo Uomo applies the same logic even more explicitly. Independent since 2022, it runs on subscriptions, keeps most of its articles and podcasts open, and treats exclusive content as a form of recognition for the people who make the project possible: a narrower audience, a smaller scale, a more tightly knit community.

Supporting a project, though, is a decision people revisit, and Defector shows how fragile that support can be. The sports outlet, owned by its own staff, publishes a breakdown of its finances every year, with ballpark figures flagged as such, putting numbers out in the open that usually stay private. In the fiscal year running from September 2024 to August 2025, subscriptions brought in about $3.8 million, flat compared to the year before, while podcasts, events, merchandise, and sponsorships added another $850,000. Total revenue came in around $4.65 million, against operating costs of about $4.55 million.

Retaining existing subscribers works, the staff writes, while signing up new ones missed projections in most months. The basic math is simple – recurring costs divided by average net revenue per subscriber – and it holds up as long as no one leaves: with a thousand subscribers and a hypothetical 3% monthly churn, you need thirty new ones every month just to stay flat. The $850,000 from other activities tells the rest of the story: podcasts, events, and merchandise come from the same staff, the same audience, and the same brand, which is exactly why they hold up.

In all three cases, the free part matters more than the paid one. That’s where people decide whether your work is worth paying for, and cutting it back too early means draining the pool next year’s paying subscribers are supposed to come from – a pool that fills up slowly. The Post, after all, had ten years of publishing behind it before it launched subscriptions.

Falsissimo sells the chance to know something before everyone else

Building trust takes time, and some people have found a way to get paid right away by selling something completely different: being the first to know.

Falsissimo, Fabrizio Corona’s show, posts a free segment on YouTube that draws a big audience, teases a reveal, and saves the payoff for channel subscribers. Here, the free content works as a trailer rather than proof of reliability, creating urgency that drives people to pay within minutes.

Setting aside the ethical and legal questions around the “Corona method,” the model can be replicated: what’s for sale is early access to information presented as exclusive. The scarcity is real, because a given system can only pull from what’s already been published, and it lasts only a few hours, until someone else picks it up. It’s the opposite of what keeps Il Post running: there, time strengthens the relationship; here, it burns through it, because each episode has to top the last one, in front of an audience that gets used to the reveal fast and expects an even bigger one next time.

The Six Types of Paywall and What Each One Asks For

Whatever the reason your audience pays, at some point you have to decide what to lock away.

The technical tools for doing this are generally called paywalls, a category that covers options asking very different things of readers. Picking one just because it worked for someone else is the fastest way to end up with nothing.

• Hard paywall – everything is locked, you read only if you pay. This works where the information drives costly decisions, typically in professional fields.
• Metered wall – a limited number of free articles per month, then you pay. It keeps discovery open and aims to convert repeat visitors.
• Freemium – some content types stay free, others stay locked, split by format or by depth.
• Registration wall – you ask for an account instead of money. You don’t get paid right away, but you gain the ability to identify readers within the consent they give you.
• Newsletter wall – you unlock content in exchange for a signup, and build the list that can later turn into subscriptions, sponsorships, or sales.
• Offerwall – you line up alternatives, from watching a rewarded ad to handing over an email address to paying outright. Google makes this configurable inside AdSense and Ad Manager.

The value of the exchange, though, has to be built beforehand: when you’re asking for money, people who stay outside need to understand what they’re missing; when you’re asking for data or attention, the benefit has to be clear and proportionate. And the wall should go up only once there’s something worth protecting.

Monetizing Expertise Instead of Content

Getting paid for reading remains a niche business in the Italian market: fewer than one in ten people say they’ve ever paid for anything to stay informed online. Yet the same people who won’t pay ten euros a month for your articles will happily spend three hundred on a course that solves a problem they have right now, and that opens up a whole different line of revenue, one that has little to do with reading: what you’re selling stops being content and becomes the expertise behind it.

The questions in the comments, the emails you get, the pieces that get read all the way through: a blog that’s been publishing for years constantly generates signals about what its audience would actually pay for. Skip that step and you end up building the course you want to teach, not the one people want to learn, and you find out only after launch – three months of work in, with twelve sign-ups to show for it.

Checking this costs almost nothing if you sell before you build: open a waitlist, or sell the first cohort with a set syllabus and dates, and if nobody signs up, you’ve lost a week.

Products You Build Once, Services You Deliver Every Time

Courses, reports, templates, consulting, communities, events: all of these keep the site open. They keep bringing people in, and what you sell sits one level further down the funnel. Underneath those labels, though, are economics that behave in opposite ways. A recorded course, a report, or a template, you build once and sell many times: the big cost is upfront, though updates, support, fees, and acquisition still add up and grow along with your customer base.

Consulting works differently: revenue per client is much higher, but there’s a hard ceiling set by the hours you have. If you want to grow without hiring, products you build once are the most direct path, though not the only one, since you can also grow by raising your prices, being more selective about clients, or automating parts of the service.

Where the value sits changes too. A course is worth as much as the method it teaches, and it goes stale if you stop updating it. A community, once it hits a certain critical mass, starts holding value for the people inside it too: members join partly to talk to each other, and at that point your job shifts from producing content to keeping a place alive. It’s the model with the heaviest day-to-day workload, since moderation never takes a day off, and it’s also the one that empties out fastest if participation drops.

On all of this, the free part works differently from the paywall. Open content is proof of your method, not a gift waiting for conversion. Someone who reads thirty articles on how to do something arrives at the course already convinced you know how to do it, and that conviction is worth more than any sales page. One question is still worth asking before you add anything, and it comes from the Defector lesson: how much does it reuse what you’ve already paid for? An event brings together the community you’ve built, a course turns into a product a method you’ve already shown. But when the new venture requires a different audience, skills you don’t have, and its own structure, you’ve stopped diversifying: you’re launching a second business and funding it with the blog.

When a Company Is the One Paying

All the paths covered so far ask your readers to reach for their wallets, and not every audience will. With sponsorships the customer changes: instead of the reader, there’s a company paying you to reach those people.

How much they pay depends on how you sell. Banners that appear automatically on your pages end up in an auction where your space is worth whatever any similar space is worth. The system can assess context and a few audience signals, but not the relationship you’ve built. If your readers are a recognizable group, though – dentists, hotel managers, procurement heads in a given industry – a company in that field will pay much more to reach them, and you negotiate with them directly.

What you’re selling them is no longer space, but a custom-built format: a sponsored newsletter, an industry research piece, a webinar, a podcast with a single sponsor, an editorial project that runs for months. You set the price, within a market shaped by budgets and seasonality, and you negotiate based on what your audience is worth rather than how many times an ad appeared.

There’s a limit past which the model eats itself, though. A reader who hits four commercial articles in a row stops seeing you as a source, and at that point you’ve burned through the asset you were selling. How far you can push it depends on how relevant the sponsor is to your readers and how the commercial piece fits into the content: a coherent partnership can sustain high frequency, while a promotion that doesn’t fit does damage from the first run. And if the sponsor starts dictating topics and conclusions, you collect one campaign and lose the reason you could have sold another.

The Blog That Sets Up a Sale Has Lost the Least

For an ecommerce site, a subscription software company, or a professional practice, the blog doesn’t earn money on the page itself. It works to make a problem recognizable, to get you onto the list of possible solutions, to demonstrate expertise before anyone asks for a quote, and the revenue shows up downstream as an order, a trial activation, or a conversion.

This is also where the drop in clicks behaves in the least intuitive way. You lose the informational visits, the ones that used to come in at the start of the journey, while the people arriving now do so at a different stage.

In the sample of US retail sites measured by Adobe, traffic coming from assistants converts 54 percent better than non-generative sources, with revenue per visit 53 percent higher. The same people spend 53 percent more time on the site and view 23 percent more pages. So someone arriving from an assistant still has something to evaluate: otherwise they’d buy right away and leave. They browse more, and the most reasonable reading is that they arrive with the field already narrowed – the assistant has already done the initial filtering, the part where you figure out what you’re looking for and rule out anything off-topic.

OpenAI, meanwhile, has acknowledged that the first version of Instant Checkout didn’t offer the flexibility it wanted, and has chosen to leave merchants in charge of their own payment flows, focusing instead on product discovery. The comparison happens inside the assistant, while the transaction and the customer relationship can stay on your own domain. That changes which pages actually generate revenue: generic guides still matter – they build coverage, earn links, and educate people who don’t yet know they have a problem – but interchangeable content counts for less and less at the moment of choice.

You need content that factors into the decision:

• which customer the product is a poor fit for, and what to do instead in that case
• which differences actually change the user experience
• which costs show up after purchase
• which results were achieved, with timelines and starting conditions
• which limitations remain even in the solution you’re proposing

Take an online store selling mountain gear: a guide on how to choose trekking boots reaches people who are still far from buying, and it’s the kind of content a generative answer can summarize with no trouble. A hands-on review of several models, though – with the conditions it was tested in, the miles covered, the flaws that showed up with use – produces something an assistant can cite but couldn’t have experienced itself, and its value lies precisely in being able to verify who did it and how.

Your margin target divided by margin per customer gives you the number of customers you need, and from there you work backward to contacts and visits using your conversion rates. The result has to fit within your actual operating capacity: a consultant who can take on eight clients a month needs the right leads more than a large number of them – for him, more traffic could even make things worse.

How to Pick the Right Path for Your Project

Visits serve all these models, but not in the same way. For advertising, they’re the product: what you sell is the page that opens. For a subscription, they’re the raw material, the pool from which the list that will eventually pay comes. A small blog with the right readers is worth little under the first logic and a great deal under the second.

The same goes for everything else you have. A recognizable voice makes a newsletter convert from day one, and it matters for an ecommerce site too, though there it works on a different lever: trust in the person recommending, rather than willingness to pay for reading. An audience that belongs to a specific category is worth nothing in an automated auction and worth a lot to a company in that industry, which pays for the relationship on top of the space.

The models overlap, and many running projects combine more than one, with one carrying the main weight. What changes from case to case is which one, because that’s what drives your editorial decisions: what you publish, what you keep gated, where you distribute it.

The distinctions you’ve seen throughout this article are mainly useful for ruling things out. If advertising requires a million page views and you’re getting a hundred thousand, that number just saved you twelve months of work in the wrong direction, and it also tells you where to look, because the right hundred thousand people can be worth far more to someone than they’d ever be worth in an auction.

Traffic is still the ingredient that fuels almost all of these paths. What it has lost is the automatic mechanism that turned ads and affiliate links into revenue, and that changes everything you need to build around it.