On 25 July 2025, the UK's Online Safety Act began requiring pornography sites to run "highly effective" age checks on British users. Within hours, Google searches in the UK for "porn sites" and its variants were running at nearly seven times their normal volume. Search traffic to toppornsites.com, a directory that ranks porn sites rather than hosting content, went from 4,967 clicks on 23 July to 23,852 on 25 July. Two smaller directories we operate saw the same multiple on the same day.
The demand shock was real, sharp and short. It halved within a week and had largely normalised within months. What the directories kept was something else: better rankings.
This is what sixteen months of first-party search data show.
Context — third-party figures for the same period
Pornhub's parent company Aylo told the BBC UK visits fell 77% after 25 July (TechRadar).
ProtonVPN reported a 1,400% rise in UK sign-ups on the day (The Register).
Similarweb data seen by LBC showed non-compliant sites growing up to 800% in a month (LBC).
Section 01
What happened on 25 July
Search Console records clicks and impressions by day. In the sixty days before the law took effect, toppornsites.com averaged 4,327 clicks a day from UK Google searches, at an average position of 3.5.
| toppornsites.com, UK | Clicks/day | Impressions/day | Avg. position |
|---|---|---|---|
| 60 days before (25 May – 23 Jul) | 4,327 | 18,717 | 3.5 |
| 23 July | 4,967 | 20,101 | 3.7 |
| 24 July | 10,578 | 44,663 | 2.9 |
| 25 July | 23,852 | 93,190 | 2.5 |
| 26 July | 25,017 | 92,861 | 2.7 |
| First week after | 20,501 | 73,429 | 2.8 |
| 60 days after (25 Jul – 22 Sep) | 16,026 | 63,723 | 6.6 |
The half-step on 24 July is the law, not noise: Search Console reports days in Pacific time, so its "24 July" runs until 08:00 BST on the 25th and contains the first eight hours of enforcement. The same half-step appears on all three sites.
Chart 1 — UK daily clicks, three directories
Indexed to 100 = average of the 60 days before 25 July. 15 May 2025 – 14 September 2026. Vertical line at 25 July.
toppornsites.com, UK — indexed, 100 = 60-day pre-law average
Section 02
More searches, or better rankings?
Clicks can rise because more people search or because Google ranks you higher. To separate the two, we isolated the family of queries containing "porn sites". The four main ones — porn sites, free porn sites, best porn sites, top porn sites — account for 90% of the family's clicks and 88% of its impressions, and toppornsites.com already held position 1–2 on all of them in the UK. When a page sits at the top of the results, its impressions are a close proxy for the number of searches.
| UK, queries containing "porn sites" | Impressions/day | Avg. position | Clicks/day |
|---|---|---|---|
| 60 days before | 5,614 | 2.0 | 1,658 |
| 23 July | 5,773 | 1.7 | 1,906 |
| 25 July | 39,730 | 2.0 | 10,364 |
| 26 July | 37,840 | 1.8 | 11,228 |
| First week after | 28,828 | 1.9 | 8,765 |
| 60 days after | 13,754 | 1.3 | 5,907 |
| September 2026 | 6,539 | 1.1 | 3,005 |
Position did not move on the day: 1.7 before, 2.0 after. Impressions did: ×6.9 on 25 July. British users searched for lists of porn sites almost seven times as often as they had two days earlier. Click-through rate actually fell, from 33% to 26%, which is what you would expect from people comparing options rather than heading for a known site.
Chart 2 — Impressions and average position, "porn sites" query family, UK
Same window.
Impressions/day, UK, queries containing "porn sites" — position in the label
Section 03
The control: Germany
Any single-country jump could in principle be a Google update or a change on the site. Germany had no age-verification law in force in July 2025 and is the second-largest European market for toppornsites.com.
| toppornsites.com | 23 July | 25 July | 26 July | 60 days before → after |
|---|---|---|---|---|
| United Kingdom | 4,967 | 23,852 | 25,017 | 4,327 → 16,026 (×3.7) |
| Germany | 11,339 | 11,459 | 11,212 | 11,239 → 12,499 (+11%) |
Same site, same days, same Google. One country moved.
Chart 3 — toppornsites.com daily clicks, UK vs Germany, indexed
Same window.
United KingdomGermany (control)Both indexed, 100 = each country's 60-day pre-law average
Section 04
France: the switch
France offers something the UK does not: the law was switched on, off, and on again within six weeks.
On 4 June 2025, Aylo made Pornhub, YouPorn and RedTube unavailable in France rather than comply with the country's age-verification rules. On 16 June a Paris administrative court suspended the decree and the sites returned around 20 June. On 15 July the Conseil d'État lifted the suspension and Aylo blocked France again.
| toppornsites.com, France | Clicks/day |
|---|---|
| 15 May – 3 June (baseline) | 1,304 |
| 3 June → 4 June (Aylo blocks) | 1,324 → 2,099, then ~2,000–2,200 for sixteen days |
| 19 June → 20 June (Pornhub back) | 1,803 → 1,370, back at baseline for 25 days |
| 14 July → 15 July (Aylo blocks again) | 1,361 → 1,930 → 2,231, and stays there |
Directory traffic tracked the availability of one company's sites like a light switch, in both directions. The magnitude also fits: in France a single group went dark and the directory gained about 60%; in the UK every major site introduced checks on the same morning and the directory gained 380%.
Chart 4 — toppornsites.com daily clicks, France
15 May – 15 September 2025, with markers on 4 June, 20 June and 15 July and a shaded band for the days Pornhub was unavailable.
Pornhub unavailable in FrancePornhub availabletoppornsites.com, France, clicks/day
Section 05
Three sites, one day
The effect is not a quirk of one domain. Two other directories we operate, with UK traffic 25 to 90 times smaller, moved on the same day.
| Site (UK) | 23 July → 25 July clicks | Multiple | Avg. position 23 → 25 July |
|---|---|---|---|
| toppornsites.com | 4,967 → 23,852 | ×4.8 | 3.7 → 2.5 |
| toppornsites.co.uk | 185 → 897 | ×4.8 | 14.0 → 8.1 |
| mypornbible.com | 267 → 811 | ×3.0 | 12.7 → 10.7 |
Both smaller sites were on the second page of Google before 25 July. Demand overflowed into page two, and both moved up several positions the same day. We report the ranking movement as an observation; we did not audit it query by query.
Chart 5 — Day before, day after and 60-day average, per site and country
UK, day before → day after (bar = share of the day-after value)60-day average afterGermany control
Section 06
Fourteen months later
The demand shock faded. Searches for the "porn sites" family in the UK were 2.4× baseline after sixty days and 16% above baseline in September 2026.
Traffic, however, did not return to where it started, at least for the largest site. toppornsites.com still receives about 13,600 UK clicks a day, three times its pre-law level. The gain now comes from rankings rather than search volume: its position on the "porn sites" family improved from 2.0 to 1.1 over the period, and from autumn 2025 it began appearing for the bare query "porn" in the UK, a change we also see in other countries and therefore do not attribute to the law.
The two smaller sites did not hold their gains. toppornsites.co.uk is back at roughly its pre-law level; mypornbible.com is below it. The day-one effect was identical across sites. What each site kept depended on ordinary search factors.
Section 07
What the data say, and what they don't
Age verification did not reduce British demand for pornography on the day it arrived. It redirected it. For a few weeks, a large share of that demand went through Google, looking for sites that had not put up a check. Ofcom's own enforcement notices point the same way: it later prioritised investigations into sites whose UK user numbers had risen significantly since 25 July.
The data cannot say where those users ended up, whether they were adults or minors, or whether they used VPNs. They cannot measure US state laws, because Search Console reports by country only. And they describe three directories run by one operator, not the market.
Section 08
Methodology
Source: Google Search Console performance reports for the three properties, "Web" search type, last 16 months (15 May 2025 – 14 September 2026), filtered by country. The query family uses a "contains: porn sites" filter; it matches several hundred long-tail variants, of which the four main queries account for 90% of clicks. Search Console dates are in Pacific time; we have not shifted them. "Before" and "after" windows are the 60 days ending 23 July and starting 25 July; 24 July is excluded as a partial day. All figures are unmodified Search Console exports. One phenomenon is deliberately left out of the analysis: a site-wide increase in impressions on the bare query "porn" from October 2025, visible in every country including Germany, which is a ranking event unrelated to any law.
MORE ANALYSIS
