Ask someone under thirty how they found the last piece of clothing they bought and the answer is increasingly not "I searched for it." It is "it came up." The Business of Fashion's work with TikTok on cultural relevance describes an environment where discovery, research and purchase now collapse into a single scrolling session — and where the platform functions less like a feed and more like a search engine with opinions.
Traditional search still carries the majority of query traffic; industry analysis for 2026 puts search engines at roughly 80% of search activity. But the first exposure to a product, and the trust that decides whether it converts, has moved. Understanding that split is the difference between shopping well online and being merchandised at.
From feed to search bar
The behavioural shift is specific: people type product-shaped queries into social apps. Not "summer outfit inspo" but "linen midi dress tall girl", "co-ord set that isn't see through", "flat sandals wide feet". These are search queries with fit anxiety baked in — the questions shoppers do not trust a brand's own product page to answer.
That is why the content that ranks in social search looks nothing like advertising. It is a person of a stated height and size, in a stated size of the garment, moving. Static campaign imagery cannot compete with it, because it cannot answer the underlying question, which is almost always: what will this look like on a body like mine?
Social search is not asking to be entertained. It is asking to be reassured.
What "cultural relevance" now costs
The uncomfortable finding in the BoF × TikTok research is that relevance is not purchasable in the way media impressions are. Platforms reward participation in an existing conversation, and the half-life of any given conversation is short. Brands that treat the platform as a broadcast channel get reach without recall; brands that treat it as a place to answer questions get recall without needing much reach.
Meanwhile TikTok Shop's sales growth has pulled the transaction itself inside the app, which changes the economics again. When discovery, review-reading and checkout happen in one place, the traditional website's job narrows to a specific and unglamorous function: being the source of truth. Detailed measurements, honest fabric information, real return terms.
This is where a lot of brands are caught out. A viral video can generate demand that the product page cannot support — and the returns rate tells the story a month later.
AI shopping agents enter the funnel
The second structural change is the arrival of AI assistants as a research layer. The State of Fashion 2026 flags the AI-assisted shopper as a defining theme, and the mechanics are worth knowing because they are unforgiving.
An assistant asked to compare three midi dresses does not look at photography. It reads text: title, description, specifications, structured data, reviews. Whatever is expressed only as an image — the actual length, the opacity of the fabric, how the waist sits — is invisible to it. Products described in mood words get summarised into mood words. Products described in numbers get recommended with confidence.
Practically, that means the brands that will surface accurately in AI-mediated shopping are the ones publishing:
- Garment measurements per size, in the units their market uses — inches for US shoppers
- Fabric composition and weight, not just "soft" or "premium"
- Fit intent stated plainly: true to size, size up for an oversized look, runs short in the body
- Care and construction detail that a returns-averse shopper would ask about
We publish inch-based chest, waist, hip and length figures on every size for exactly this reason, and it is why our New In pages lead with specification rather than adjectives.
How to read a viral product before you buy
A short protocol that filters most disappointment out of social-discovered purchases.
Find the creator's stated measurements, then find the garment's. A dress that hits mid-calf on someone 5'4" hits above the knee on someone 5'9". If the creator does not state their height and the size worn, the video is decoration, not information.
Check the fabric against the movement. Watch how the garment behaves when the wearer turns. Fabric that swings and returns has weight; fabric that clings and creases in the same shot is light and will show everything.
Search the product name plus "return". The most useful reviews are written by people who sent something back, because they explain precisely which dimension was wrong.
Assume the colour is one shade off. Ring lights and app filters warm every neutral. If the exact tone matters — because you are matching it to something you own — buy from a brand that publishes a flat, unfiltered product shot as well as an on-body one.
Verify the fit intent, not the size. "Oversized" from one label is a straight fit from another. The garment's measurements versus a piece you already own is the only apples-to-apples test. Coordinated pieces are the easiest place to start, since our two-piece sets are graded to be worn separately, which gives you two chances to get a wear out of each half.
Frequently asked questions
Do people really use TikTok as a search engine for fashion?
Increasingly, yes — particularly for fit- and occasion-specific queries where shoppers want to see a garment on a real body. Conventional search engines still handle around 80% of overall search traffic, so the honest picture is a split funnel: social for first discovery and reassurance, search engines for comparison and purchase intent.
Why do clothes look different in videos than on the product page?
Three reasons: lighting and in-app filters shift colour temperature, motion flatters drape in a way static images do not, and the creator's height and size change where every hemline falls. Compare stated measurements rather than impressions.
How do AI shopping assistants decide what to recommend?
They read text and structured data — titles, descriptions, specifications, schema markup and reviews. Attributes shown only in photographs are effectively invisible, which is why precise written specifications now materially affect whether a product gets recommended.
What should I check before buying something I saw on social media?
The creator's height and the size they are wearing, the garment's own measurements in inches, fabric composition and weight, the return window, and at least one review from someone who returned it. That sequence removes most fit-related surprises.
Sources
- Building Cultural Relevance on TikTok, The Business of Fashion × TikTok — businessoffashion.com/reports
- The State of Fashion 2026, The Business of Fashion & McKinsey & Company — mckinsey.com
Editor's note: research and drafting for this article were AI-assisted and reviewed by the PHEFEE editorial desk before publication.
