Why generated adult sites don't rank, and what changes when every page has its own text

Every few months someone launches a site generator for adult. Pick a niche, click, get a tube with ten thousand videos. Two months later the site has a handful of visits a day and most of them are the owner checking it. The generator didn't lie, it built a site. It built one Google had already seen thousands of times.

We build one of those generators, so take this with that in mind. We wrote about it on the blog in Why generated adult sites do not rank. This version goes further into the part people argue about most, what "unique text" really buys you and what it doesn't.

The pool problem

Adult video catalogs are shared. The same tube networks feed thousands of affiliate sites, and every video comes with the same title and the same two lines of description. A site that serves them as they arrive is a copy of every other site on that network, page for page.

Google has two names in its spam policies that fit this situation almost too well. "Thin affiliation" covers pages with affiliate links where the descriptions are "copied directly from the original merchant without any original content or added value". And "scaled content abuse" is when "many pages are generated for the primary purpose of manipulating search rankings and not helping users". A feed poured into a template with no changes is the first. Ten thousand pages of it on day one starts to look like the second.

It's not a penalty in most cases, and that's what makes it confusing. Nobody gets a manual action email. The pages simply aren't worth indexing when an older, stronger site already carries the same text, so Google picks that one and leaves yours in "Crawled, currently not indexed", where you find a few thousand of them a month later.

What unique text buys

Rewriting every title and description for your site fixes the duplicate problem at its root. Each video page now carries text that exists nowhere else, so Google has no older copy to prefer. On a tube, where the page is mostly a player and a grid, those two fields are most of what a crawler reads, so changing them changes the page more than you'd think.

It also lets you write for your niche. A feed title is written for every site at once, often keyword-stuffed or just a studio's scene code. A rewritten title can use the words people in your niche actually search for, the way they type them. On a site about one narrow niche, that's the difference between ranking for its terms and ranking for nothing.

What unique text doesn't buy

Here's the uncomfortable part for anyone selling rewriting, us included. Google's scaled content policy explicitly covers pages made with generative AI "without adding value for users". Replacing a feed's text with a model's text removes the duplicate, but by itself it doesn't make a site worth ranking. Ten thousand rewritten descriptions dumped on a new domain in one night are still ten thousand pages nobody asked for.

So the text has to come with the rest of what a real site has. Three things matter most in our experience.

A pace a crawler can follow. A new domain gets a small crawl budget. Publish past it and the extra pages wait unindexed, and a site that's mostly unindexed pages gets judged on that. We went into the arithmetic in Ten thousand pages on day one. The short version is to launch with a modest batch dated over the previous weeks, then add a steady number of pages every day, each with its own date.

Structure beyond the video pages. A site that's only a list of videos has nothing that explains what it is. Niche pages, a few real articles about the category with their own titles, linking to the videos and linked from the navigation, give the site a spine that search engines read as topical depth. Performer pages do something similar for names people already type.

Freshness that doesn't stop. A site that got its ten thousand pages and never another one gets visited less and less. The catalog behind it keeps growing, so the site should keep publishing from it. You don't need a lot every day. You need something new on most visits.

None of those is exotic. They're what a person running a site by hand does without thinking. Generated sites fail because they skip them, and they skip them because the whole pitch is not having to think.

A concrete comparison

Take two sites on the same niche, same catalog, same template, and what we'd expect to see from each.

The first launches with eight thousand videos, feed titles, no other pages. A few weeks in, Search Console would typically show a few hundred pages indexed and thousands "discovered, currently not indexed". The indexed ones rank on long-tail scene titles that a hundred other sites also rank on, somewhere on page four.

The second launches with sixty videos dated over three months, each with its own title and description, and a niche page explaining the category. It adds six videos a day. After the same few weeks it has fewer pages in total and a far larger share of them indexed, and the niche page starts picking up impressions on the category term itself.

There are no traffic numbers here for either, because they depend on the niche, the domain and luck, and we won't invent them. What we can say is which of the two Google has a reason to keep crawling.

What we'd do on any stack

If you build your own sites, with a script, WordPress or anything else, the rules are the same.

Don't publish the feed's text. Rewrite it, by hand or with a model you brief properly, and read a sample of what comes out, because models write very confident nonsense about videos they can't see. Keep the search terms of the niche as people type them, even the ugly ones.

Launching everything at once is the other classic mistake. Set a publishing calendar and stick to it. Backdating a first batch is fine as long as the dates are plausible.

Don't unpublish pages that are indexed. When a site's budget changes, stop the new pages and leave the old ones. Taking down indexed pages hands back rankings that took months to earn.

Add pages that aren't videos. One good niche page is worth more than a thousand more video pages on a young domain.

Pick a niche narrow enough to be about something. A dozen categories make a general tube, and general tubes are the one kind of site where the big ones will always win.

How Jimp handles it

Disclosure first, we build this. On a Jimp site, rewriting isn't an option you can skip, every catalog video gets its own title and description before it goes online, and it costs no credits. The pace is fixed per site type and the same on every plan, sixty videos over ninety days at launch, then six a day. Niche pages exist, written by you or generated with credits. Nothing indexed is ever unpublished when a plan changes. The publishing and SEO page has the details.

We made those things compulsory because a generator that lets you skip them will be used that way by someone in a hurry, and the sites will fail. If you think we've got the pace wrong for your niche, or you've seen different results with a different approach, say so here with what you measured. That's the discussion we'd like this category to have.