Paywall or Ads? Revenue Model and Google Paywall Markup

Short answer: there is no universally correct model. Ads suit a site with broad search traffic and readers who rarely return. A subscription suits a publication with a loyal audience and content that cannot easily be found elsewhere. A metered paywall suits the large middle ground. Whatever you choose, a paywall brings a second, technical decision: if Google cannot tell that hidden content is deliberately gated, your pages can look like they show one thing to visitors and another to the crawler.

This guide is written for niche publishers and the owners of news and guide sites. It compares three revenue models, gives a blank worksheet for the revenue arithmetic, explains how Google's subscription and paywalled content markup works and where its spam policy draws the cloaking line, covers in-between options such as a registration wall and a newsletter, and closes with a way to run the switch as a controlled test. It is not legal advice, but access and markup decisions can affect how your pages appear in search, so apply them carefully.

Three Models and Who Each Suits

A revenue model is a statement about what you are selling. Advertising sells attention. A subscription sells the content itself, or the relationship around it. A metered paywall sells a little of both, and decides who gets which by counting. Each choice changes your editorial priorities, your technical setup and what a search visit is worth.

Ads only

Everything is public and revenue comes from ad impressions and clicks. A reader arriving from a search result, a social post or a shared link sees the article immediately, which makes this the easiest way to reach new readers. The trade-off is that revenue tracks traffic volume and the value of your ad inventory. Readers who use ad blockers add nothing, a problem covered in ad blockers and publisher revenue, and heavy ad placements can hurt page speed and layout stability.

Metered paywall

Readers get a set number of free articles, and a subscription prompt appears once the quota is used. Google's flexible sampling page describes two approaches: metering, which gives users a quota of free articles before asking for a subscription or login, and lead-in, which shows part of an article and withholds the rest. A meter lets a first-time visitor from search read something, then asks for money only from people who keep coming back. It can also carry ad revenue and subscription revenue in the same product.

Full subscription

Most of the main content is unavailable without logging in or paying. Revenue is steadier and does not swing with ad markets. The cost is discoverability: someone who clicks a shared link may not see anything worth sharing. This model works when the publication offers something readers cannot get elsewhere, such as original data, deep analysis, expert commentary or a community.

Side-by-side summary

Decision rule: which one?

Work through these questions in order and stop at the first model that survives every check.

1. Can readers find a close equivalent of your content for free elsewhere? If yes, a full subscription is a poor fit.
2. Do you have a core of readers who return on their own? If not, asking for payment is premature.
3. Does most of your revenue and growth depend on search visits? If yes, keep the first visit free, which points to ads or a meter.
4. Have you already improved what ads earn per page? If not, work on metrics such as those in RPM, eCPM and fill rate first. It is cheaper and carries less search risk than building a paywall.

A reader's path under each model

Under ads, a visit is usually a single contact: arrive, read, leave. Under a meter, a returning reader reaches a decision point after several articles, and the site learns something about intent from what that reader does next. Under a full subscription, the decision happens on the first visit. The same article therefore plays a different role in each model: an advertisement for the publication in one, and the product itself in another. Decide which articles are doors and which are rooms before you decide where to put the lock.

Back-of-Envelope Revenue Comparison Sheet

You cannot know in advance which model will earn more, but a simple worksheet makes your assumptions visible and tells you afterwards which assumption was wrong. The sheet below is deliberately left blank for your own numbers.

Ad revenue chain

Monthly ad revenue is the product of sessions, pages per session, ad impressions per page and revenue per thousand impressions. Each link is a separate lever. A paywall mainly reduces pages per session, because a reader who hits the quota stops browsing. To estimate the effect, you need to know how many of your readers actually reach the quota, and that is a measurement, not a guess.

Subscription revenue chain

Monthly subscription revenue is the product of unique readers who meet the paywall, the share who open the subscription page, the share of those who complete payment, the price, and how long subscribers stay. The last term is the one most often ignored. Winning a subscriber once and keeping one month after month are different jobs, and cancellations quietly erase growth.

The break-even test

Take the share of your ad revenue that comes from pages the paywall would touch. Divide that amount by the subscription price. The result is the number of paying subscribers you need just to replace the ad revenue you give up. If that number is far larger than a realistic conversion from your actual audience, the model does not fit yet, however attractive it looks.

Example scenario: a gadget review site

Example scenario: a review site gets most of its visits from search, and most readers read a single buying guide before leaving. The owner runs the break-even test and sees that replacing the ad revenue would require converting a share of casual readers that looks unrealistic. The owner keeps buying guides open and places only the long lab-style comparisons behind a meter, then measures both groups separately.

Common mistakes in the arithmetic

• Counting only lost page views and ignoring that a metered reader may visit less often afterwards.
• Treating subscription revenue as fixed instead of subtracting expected cancellations.
• Running the trial for too short a period. Reader habits take weeks to settle.
• Closing everything. Leaving no free layer cuts off the search visits that feed the funnel.

Marking Paywalled Content Without Cloaking

The riskiest part of a paywall is getting the technical side wrong. Google's spam policies (Opens in a new tab) define cloaking as presenting different content to users and search engines with the intent to manipulate search rankings and mislead users. The same page says paywalls and content gating are not treated as cloaking when Google can see the full gated content, just like any person who has access to it, and the site follows the Flexible Sampling guidance.

What the markup does

Google's subscription and paywalled content markup (Opens in a new tab) guide explains how to use schema.org CreativeWork properties in JSON-LD to tell Google that gated content is a paywall and not cloaking. The required property is isAccessibleForFree, set to false when content sits behind a paywall or needs a subscription or registration. The recommended properties are hasPart with the type WebPageElement, a cssSelector that points to the class name you added in the HTML, and isAccessibleForFree set to false inside hasPart.

The steps are: wrap each gated section in a class name, add structured data for NewsArticle or another supported type, add isAccessibleForFree at the top level plus hasPart pointing at the class, and validate with the Rich Results Test. If you have several gated sections, use an array of hasPart objects. The guide asks you to apply the same steps to every version of the page, including AMP and non-AMP versions.

Rules to respect

• Do not nest gated sections.
• Use only class selectors (.class) in cssSelector.
• JSON-LD and microdata are both accepted.
• Supported types are CreativeWork and its subtypes, such as Article, NewsArticle, Blog, Review, Course, HowTo and WebPage.
• If content must never reach the browser, pick a paywall implementation that does not deliver it.

The markup applies only to content you want crawled and indexed. If you do not want gated content indexed, the guide does not apply. It also says Googlebot, and Googlebot-News for news content, must be able to access the page.

Markup is a label, not a shield

The markup tells Google that a section is gated. It does not replace the duty to serve the same content to users and to the crawler. The guide notes that violating the structured data guidelines can make a page ineligible for special appearance in results. If a structured data manual action applies, the markup is ignored, although the page can still appear in search.

Client-side paywalls

Many sites build a paywall in the browser: the full text arrives, a script opens an overlay and hides it. In that setup the gated text is technically present in the page source. Google's guide says that if content must not reach the browser, you should use an implementation that does not deliver it, and it points to separate guidance on JavaScript paywalls inside its documentation on fixing search-related JavaScript problems. The practical question is whether you mind if a determined visitor can read the gated text from the source. If you do, gate it on the server.

Previews and snippets

The guide states that AI Overviews and AI Mode show previews drawn from web sources and are subject to Search's preview controls. To keep gated sections out of snippets, it names the data-nosnippet attribute, and the max-snippet robots meta tag caps snippet length. The documentation does not quantify how these controls change visibility.

Testing and monitoring

The Rich Results Test validates syntax but cannot detect guideline or spam issues. The URL Inspection tool shows how Google crawls and renders a URL. In Search Console, watch the structured data reports and the Manual Actions report. The guide says recrawling can take several days after you publish and that Google does not guarantee structured data features will appear. Google's documentation can change, so check the current pages before you implement.

Pre-launch checklist

1. The gated section has a class name in the HTML and the same class appears in cssSelector.
2. isAccessibleForFree is false at the top level and inside hasPart.
3. Googlebot can reach the page; neither robots.txt nor a login wall blocks it.
4. Googlebot and a real subscriber see the same gated content.
5. The Rich Results Test shows no critical errors.
6. The Manual Actions report is clean.

Hybrid Options: Registration Wall and Newsletter

Between full ads and full subscription there is a wide spectrum, and many publications are best placed somewhere in the middle.

What Google suggests for metering

The flexible sampling page (Opens in a new tab) recommends monthly metering over daily metering, saying it offers more flexibility and a safer environment for testing, and that a monthly limit concentrates paywall exposure on the most engaged users. It also calls showing the opening sentences after the meter runs out a good practice, because it lets users sample the value of the content.

On numbers, the page says that for most daily news publishers the optimal range is 6 to 10 articles per user per month, and as a starting point suggests giving Google Search users 10 articles per month, then adjusting. It adds that user satisfaction declines significantly when paywalls appear more than 10% of the time, which it equates to about 3% of the audience hitting the paywall. These figures are Google's expectations for daily news publishers. They may not fit your niche and can change, so check the current page and then tune against your own data.

A registration wall

A registration wall asks for a free account, not a payment. It earns nothing directly, but it turns an anonymous visit into a relationship. That relationship supports a newsletter and, later, a subscription offer. Google's markup guide applies the same approach to sites that require registration for indexed content, so a registration-gated section should be labelled the same way. If you collect personal data, review your notice and consent obligations separately.

A newsletter first

A newsletter lets you build a returning audience without a paywall. Revenue can come from sponsorship or a separate paid edition. It also loosens your dependence on search traffic by giving you a direct channel. If you are not ready to choose a model, building the list is a low-regret preparation for any of them.

Splitting a catalogue by content type

Some publishers keep news, short explainers and evergreen guides open, and reserve datasets, tools, archives and long investigations for members. Open pages bring in search traffic and show what the publication can do. Members-only pages give regular readers a reason to pay. The weak spot is a free layer that is not worth reading, because a visitor who finds no value in it never sees the offer.

Signals that you are losing readers

• Exit rate and return frequency for readers who met the paywall.
• The share of first-time search visitors who hit the paywall in their first session.
• Newsletter sign-up rate and open behaviour.
• Time on page for visits from shared links.
• The type of content last read by people who cancel.

Running the Switch as a Test

Switching the whole site at once is hard to undo. A small, measurable trial limits both revenue and search-visibility risk.

Choose a slice

Pick one content group, such as a single category or content type, and know how much traffic and revenue it contributes to the whole site. That caps the possible loss. Keep a similar group without a paywall as a comparison.

Set up and verify

1. Add the markup to the template of the gated pages.
2. Confirm that Googlebot and a real user get the same content.
3. Instrument events: paywall view, click to the subscription page, registration, payment, cancellation.
4. Check with the Rich Results Test and URL Inspection.
5. Make sure ad slots do not stack on top of the paywall prompt. Showing a heavy ad load and a paywall together damages the experience.

Write the success criteria first

Decide how long the test runs and what counts as success before you start. Otherwise the criteria drift once results arrive. Split them into revenue measures (net revenue, remaining ad revenue, subscriber count) and health measures (search traffic, share of returning readers, newsletter sign-ups). If revenue rises while search traffic falls sharply, ask whether the gain can last.

Watch search performance

In the Search Console Performance report, compare clicks, impressions, click-through rate and average position for the test group against the comparison group. Google's report help notes that the most recent data can be preliminary and may change over the next few hours, so avoid decisions based on one day's movement.

Plan the rollback

Before launch, define the condition that triggers a rollback and the steps to carry it out. Remember the markup: if content becomes free again, isAccessibleForFree must not stay false. Stale markup misdescribes the page.

Reading the outcome

Example scenario: a regional news site

Example scenario: a regional outlet has a faithful readership and unique coverage of local government. It meters its investigations and archive, leaves daily news open, marks the metered sections, and compares search clicks for the two groups after several weeks. Daily news keeps feeding the funnel while the deeper coverage tests willingness to pay.

Mistakes to avoid

• Launching a paywall without markup: Google cannot tell gating from cloaking.
• Treating the crawler differently by intent: cloaking is defined by the intent to manipulate rankings and mislead users. Stay inside the documented markup and flexible sampling guidance.
• Closing the whole catalogue at once: it narrows the path from search to new readers.
• Copying a quota from someone else: Google's published figures describe daily news publishers. Measure your own audience first.
• Forgetting to remove markup: update the JSON-LD whenever access changes.

Related guides: Google Ad Manager for Small Publishers: Setup Guide, AdSense Auto Ads vs Manual Ad Units, Do You Need a CMP on an Ad-Supported Site? How to Pick One.

Conclusion

Paywall or ads has no single answer. A publication with original, hard-to-find content and a returning audience can move toward subscriptions. A site built on general information and search traffic is safer with ads and a meter. In-between steps, namely a registration wall and a newsletter, are the lowest-risk way to build a reader relationship while you decide. Whatever you pick, mark gated sections with isAccessibleForFree, give Googlebot and users the same content, and run the switch as a small test with written criteria. Google's documentation and its thresholds can change, so check the current pages before you implement. For a wider comparison of ways to earn from a site, read how to monetize a website, and for the general logic of markup see the JSON-LD structured data guide.

If you are looking for placements to run ads on your own audience's terms, you can browse ad spaces on Akua Adforce; scope and terms are listed on the product page.

Frequently Asked Questions

Can Google index content that sits behind a paywall?

According to Google's subscription and paywalled content guide, if you want gated content crawled and indexed, Googlebot must be able to access the page and the gated section should be marked with isAccessibleForFree set to false. If you do not want it indexed, the guide does not apply. Nothing guarantees how it will appear in results.

Is showing Googlebot the full text and users a gated version cloaking?

The spam policy defines cloaking as presenting different content to users and search engines with the intent to manipulate rankings and mislead users. Paywalls and gating are not treated as cloaking when Google can see the full gated content like any person with access and the site follows the flexible sampling guidance. Mark the gated section correctly.

How many free articles should a metered paywall allow?

Google's flexible sampling page expects 6 to 10 articles per user per month to be optimal for most daily news publishers and suggests starting with 10 per month for Google Search users. Those figures are about daily news publishers and can change, so tune the quota against your own measurements.

Can I run ads and a paywall on the same page?

Yes. In a metered model free content can earn from ads and gated content from subscriptions. Avoid stacking a heavy ad load on top of the paywall prompt, and mark the gated section.

Should I remove the markup if I drop the paywall?

Yes. If the content is open again, isAccessibleForFree must not stay false. The markup should describe the page's real access state, and updates show once Google recrawls the page.

Does a free registration wall need paywall markup too?

Google's guide covers sites that offer subscription-based access or require registration for indexed content. It therefore makes sense to label a registration-gated section the same way.