Intent-Driven Category Pages: How ‘Airport Outfits’ Outsold ‘Jeans’ by 841%

In a 2026 talk on mixed SEO, AEO, and GEO trends, German search analyst Malte Landwehr reported a number that made the rest of the room sit up. PrettyLittleThing replaced traditional product-noun category pages with intent-led ones. A page called “jeans” became a page called “airport outfits.” The reported lift was 841%. Every product featured on the intent-led pages sold out of stock.

The number is striking on its own. The mechanic behind it is the part marketers should care about more.

Intent-driven category pages are the next stage of ecommerce information architecture. They don’t replace the old category pages so much as they reveal what those pages were always missing: a connection to the actual job the customer is trying to do.

A quick note on the data. PrettyLittleThing is part of Boohoo Group, which went private in 2025, so there’s no annual report or earnings call to verify the figure independently. The 841% number traces to Landwehr’s coverage. I treat it the way trade press treats any private-company stat: source it clearly, don’t dress it up as a primary disclosure.

Key Takeaways

• Landwehr reported PrettyLittleThing saw an 841% uplift after replacing product-noun categories with intent-driven ones, with every featured item selling out
• The shift moves the organizing principle of a category page from what the product is to what the customer is trying to do
• Intent pages match the longer, conversational queries users now run and the fan-out sub-queries LLMs expand a single prompt into
• The migration is mostly an information architecture exercise, not a content rewrite, and intent pages should sit alongside existing category pages rather than replace them

The Result PrettyLittleThing Reported

The headline figure: an 841% uplift, with every item on the intent-led pages going out of stock. Landwehr surfaced the case study as evidence that intent-driven category architecture beats keyword-driven category architecture at scale.

That’s the entire public record. Boohoo Group does not file financial disclosures any more, and PrettyLittleThing’s owned channels haven’t independently confirmed the figure. So the right way to read it is as a directional signal from someone deep enough in the European ecommerce SEO scene to be worth listening to, not as an audited number.

What’s harder to argue with is the underlying pattern. The same shift, with smaller numbers and verifiable sources, shows up across enough other retailers that it’s no longer fringe.

What “Intent-Driven Category Page” Actually Means

Most category pages on most ecommerce sites are organized around a noun. /jeans. /running-shoes. /sofas. /headphones. The page lists every product in that category, with filters for size, color, price, and brand. The visitor is asked to do the work.

An intent-driven category page is organized around a customer scenario instead. The same site might keep /jeans, but it adds /airport-outfits, /work-from-home-fits, /first-day-of-school. Each of those new pages curates a specific set of products that solve a specific problem, often pulling from multiple traditional categories at once.

Here’s the difference side by side.

Other retailers running variants of the pattern: REI’s “first backpacking trip” guides, Lululemon’s “yoga to brunch” collections, Wirecutter’s scenario-based gear roundups, and B&H Photo’s gear-by-use-case pages. The execution differs but the principle is the same. The page itself is the recommendation, not the directory.

Why Intent-Led Pages Win on Modern Search

Several shifts in search behaviour have made intent-led pages the better fit for how users actually search. The pattern is not new, but AI search has made it stick.

Queries got longer

The Q4 2025 State of Search data showed average query length climbing across the board, with conversational, multi-clause queries becoming routine. I covered this in detail in Search Queries Are Getting Longer. A page named after a single noun cannot match a query that names a scenario.

Users describe problems, not products

Ask any merchandiser what their site search log looks like and they’ll tell you the same thing. People type “jeans for tall guys with short waists” or “comfortable shoes for standing all day,” not “jeans” or “shoes.” Site search has been showing us the gap for years. Most category architectures haven’t caught up.

LLMs fan out prompts into intent sub-queries

When someone asks an AI assistant “what should I wear on an early flight,” the model expands that into a fan-out set: comfortable layers, easy-to-remove shoes, breathable fabrics, no underwire, slip-on accessories that pass through TSA without drama. An intent-led page on /airport-outfits answers most of that fan-out set on one URL. A keyword-led /jeans page answers none of it.

The first two patterns have been building for a decade. The third one changes the math fast.

Why Keyword-Led Pages Lose Conversions

The conversion gap is at least as big as the traffic gap, and the cause is straightforward.

A /jeans page hands the customer a problem disguised as a catalog. They have to filter by fit, then wash, then length, then color. They have to know which brand suits their shape. They have to scroll past hundreds of products that aren’t right for them. The bounce rate on a generic category page reflects the cognitive load of all that, and the AOV reflects the fact that people who do convert usually buy one thing.

An /airport-outfits page does the editing. It says: this top, with these joggers, with these slip-ons, plus a layer for the cabin. The visitor’s job is to confirm or replace one item, not to compose an outfit from a thousand SKUs. The page sells four items at once because the page is itself the cross-sell.

You can see the math working in the PrettyLittleThing report. Sell-through wasn’t just up. It went to zero remaining stock. That kind of clean-out usually comes from one of two things: a viral product everyone wants, or a curated bundle that removes the choice paralysis that keeps people in browse mode.

How to Identify Intent Categories Worth Building

Most ecommerce teams assume they need a sweeping IA project to do this work. They don’t. The signal is already in their own systems.

Mine the questions, not the keywords. Six places to look:

1. On-site search logs, especially the queries that return zero or weak results
2. Customer service tickets, particularly recurring “what should I get for…” questions
3. Reddit threads in your category’s main subs, where people ask scenario questions
4. LLM prompts that mention your category, accessible through tools that aggregate them
5. Your reviews section, where buyers describe the use case they bought for
6. Your competitors’ breadcrumb structure, which sometimes reveals intent pages they’ve already validated

For each candidate intent, validate two things. First, is there real demand? If the long-tail variants of the intent phrase together pull more search volume than the head term, the intent category is real. Second, can you actually merchandise it? An intent page only works if you have or can curate the products to fulfil the promise.

The Migration Plan Without Burning Existing Rankings

The biggest mistake teams make when they hear about intent-driven IA is replacing their keyword-led pages. Don’t. /jeans still ranks, still gets organic traffic, and still converts the searchers who type “jeans.” Killing it gives up real revenue to chase a thesis.

Six steps that move you toward intent IA without breaking what already works.

1. Keep every existing keyword-led category page. None of them get redirected.
2. Build the new intent pages alongside, with their own URLs, breadcrumbs, and templates.
3. Cross-link both directions. The /jeans page surfaces /airport-outfits and /work-from-home-fits in a “wear them for” module. The /airport-outfits page surfaces /jeans, /joggers, and /loungewear in a “shop the category” module.
4. Add intent pages to the main navigation as a second axis. Keep the product-type axis intact and add an “Outfits for…” or “Shop by occasion” axis next to it.
5. Watch for cannibalization in Google Search Console at the URL level for the first 60 days. If the intent page starts to absorb queries from the keyword page, that’s the system working correctly. If both pages compete on the head term, fix the targeting.
6. Track AOV and add-to-cart rate per page, not just clicks. Intent pages often look weaker on traffic and stronger on revenue. The traffic-only view will mislead you.

That’s enough structure to start. You don’t need a six-month IA project. You need one or two intent pages, shipped, measured, and iterated on.

What This Means for Your AI Visibility

The case for intent-led category pages used to rest entirely on conversion. The AI search shift adds a second case that’s nearly as strong: citation.

LLMs cite passages, not pages. A clean recommendation block at the top of /airport-outfits (“for a 6am flight, layer a soft long-sleeve over a fitted tank, joggers with a stretchy waistband, and slip-on sneakers without metal eyelets”) is exactly the kind of self-contained chunk an AI assistant lifts when someone asks the question. The same prompt against a /jeans page returns nothing usable.

I covered the citation mechanics in more depth in How to Get Cited by ChatGPT and in the Google AI Mode SEO guide. The pattern there is the same as the conversion pattern here. Pages that do the curation get rewarded. Pages that hand the work back to the visitor or the model get skipped.

It’s also worth keeping the broader click-economy context in view. As I showed in Zero-Click Searches: The Numbers Behind the Panic, the click rate on traditional search has stayed steadier than the alarmist takes suggest. But the clicks that do happen are increasingly going to pages that look like answers. Intent-led category pages look like answers. Keyword-led ones look like phone books.

FAQ

Sources & References

1. Malte Landwehr (2026). “Mixed SEO, AEO, GEO takeaways.” X (formerly Twitter). https://x.com/MalteLandwehr/status/2050268682115272994
2. Datos and SparkToro (2025). “State of Search Q4 2025.” https://datos.live
3. REI scenario-based gear pages. “First Backpacking Trip” guide series. https://www.rei.com/learn/expert-advice
4. Wirecutter (2026). Scenario-based recommendation guides. https://www.nytimes.com/wirecutter/
5. Lululemon “Yoga to Brunch” collections. https://shop.lululemon.com/