what actually changed about tube design in five years

the layout changed first, then the incentives followed. that’s the order most people get backwards when they talk about tube design.

mobile took over and rewrote the whole layout logic, not just the css

somewhere between 2019 and now, adult traffic went overwhelmingly mobile, something like 87% of sessions on most major tubes. that’s not a stat you can patch around with a responsive breakpoint. it forces a different information architecture entirely.

the old desktop grid, 4 or 5 thumbnails across with a category rail on the left, basically died. what replaced it is a single-column vertical feed you scroll through with your thumb, closer to instagram reels than to a video library. pornhub’s redesign history over the last five years traces this exact arc: fewer visible categories per screen, more autoplay preview on hover-or-tap, thumbnails getting taller and more cropped to fit a phone screen instead of a desktop monitor.

the tradeoff nobody advertises: category taxonomy got buried. what used to be a visible horizontal bar of 30+ categories is now a hamburger menu or a “more” button two taps deep. mobile-first didn’t just change how things look, it changed what’s discoverable by default, and category browsing lost that fight to the infinite scroll feed.

the ad model got sneakier even as the design looks cleaner

this is the part that actually annoys me. the visual noise went down. the popup walls and flashing banner grids that defined 2015-era tube design are mostly gone, at least on the big five (pornhub, xvideos, xhamster, spankbang, eporner). that reads as progress if you only look at the top of the page.

but the ad surface didn’t shrink, it just got better at disguising itself. native ad slots now sit inside the same card grid as real scenes, styled identically, so you can’t tell what’s a scrape of an actual upload and what’s a sponsored redirect until you click it. xhamster in particular has gotten good at blending “related video” rows with paid placement. spankbang leans on premium-nag interstitials that interrupt playback right as you’re 20 seconds in, framed as a UX feature (“upgrade for uninterrupted playback”) rather than what it is, which is an ad you can pay to remove.

that’s the actual shift in tube ad walls: from obvious-and-ugly to invisible-and-embedded. dark patterns didn’t disappear when the design got minimalist, they just moved from the banner slot into the content layer itself, where you have to click to find out it wasn’t content.

creator carousels ate the real estate that used to belong to categories

onlyfans and fansly started pulling meaningful traffic and revenue away from the tube model around 2020, and the tubes noticed. the response wasn’t better categories, it was creator carousels bolted onto the homepage, usually above the fold, usually in the exact spot category tiles used to occupy.

the reason is simple and not even hidden: tubes take an affiliate cut when you sign up through a creator carousel. a category click monetizes indirectly through ad impressions. a creator signup monetizes directly and immediately. given that choice, homepage real estate goes to the thing with the clearer payout, every time.

i’ll say the unpopular part directly: this isn’t a neutral design evolution, it’s a business model creeping into product decisions at the expense of the thing tubes were actually good at, which was letting you browse by what you wanted instead of by who’s paying for placement. category discovery got measurably worse because something more profitable needed the space.

what actually improved: search, autoplay defaults, and deep-linking

credit where it’s due, because not everything went backwards. search got genuinely better across most of these platforms. autocomplete works, tag filtering is more granular, and some sites (eporner and xvideos both do this reasonably well) got smarter about surfacing results instead of just keyword-matching titles.

autoplay-into-next got both smarter and more controllable. a few years ago it was an aggressive default with no visible toggle. now most tubes let you kill it in one tap, and the algorithm behind “next” is at least loosely related to what you were watching instead of whatever’s trending globally.

deep-linking quietly improved too. scene URLs are more stable, og tags render properly for previews, and that matters more than it sounds like it should. it’s the difference between a link that works when shared and one that 404s six months later. it’s also the entire reason an aggregator model works at all, since the whole premise depends on the destination link staying valid.

the part nobody talks about: age gates and consent walls as friction by design

here’s the one that gets almost no coverage. EU pressure, the UK’s online safety act, france’s loi SREN, germany’s JuSchG, italy’s decreto caivano, spain’s RD 1/2024, all pushed tubes toward some form of age verification or geo-blocking. that’s a real regulatory shift and a lot of it is legitimate compliance work, not a conspiracy.

but watch what happens at the gate itself. some of these walls take one click. others make you click through a warning, then a region confirmation, then a “remember me” prompt that doesn’t actually remember you next session, and somewhere in that sequence there’s usually an ad slot that wasn’t there in 2020. the friction is dressed as consent and safety, and some of it genuinely is. some of it is also just an extra impression squeezed out of a screen you’re forced to look at before you can leave.

nobody writes about this because it’s boring and because criticizing age gates sounds like criticizing safety, which isn’t the point. the point is that “compliance friction” and “monetized friction” look identical from the user side, and tubes have every incentive to blur that line rather than clarify it.

that’s the honest tally after five years: mobile forced a real architectural rewrite, the ad model got quieter but not smaller, categories lost ground to creator monetization, search and linking actually got better, and the gates you click through before any of it now carry more than they claim to. if you’re trying to browse by category instead of by whatever the homepage wants to sell you this week, that’s most of why it feels harder than it used to, prowl exists because that gap is real and nobody’s mapping around it.

the layout got prettier. the incentives underneath it didn’t get any more honest.