We Audited 50 DTC Brands for AI Visibility. Social Fame Didn't Win.

A growing share of online shoppers now ask ChatGPT, Claude, or Gemini "what's the best mattress" or "is Glossier worth it" before they ever open a retailer's site. So we ran AI visibility audits on 50 direct-to-consumer brands across five categories — mattresses, sneakers, skincare, luggage, and activewear — to see who actually gets recommended.

The most surprising result wasn't about a hidden winner. It was about who didn't win: some of the most culturally famous, Instagram-dominant brands in DTC scored worse than competitors most people have never heard of.

Key Finding #1: Social Fame Doesn't Equal AI Visibility

Glossier built the modern DTC beauty playbook. Rhode, Hailey Bieber's skincare line, is one of the most talked-about beauty launches of the past few years — routinely selling out, constantly covered by fashion press, dominant on TikTok and Instagram. Both brands are, by any normal measure of fame, beauty industry royalty.

Both scored exactly 60 — a C grade — for AI visibility. Meanwhile The Ordinary, a famously unglamorous, ingredient-list-on-the-bottle skincare brand, scored 93. Ilia scored 97.

The likely explanation: AI engines aren't trained primarily on Instagram engagement or TikTok virality — they're trained on text. The Ordinary and Ilia both have dense, structured, ingredient-and-efficacy-focused content across dermatology sites, beauty review publications, and Reddit skincare communities — exactly the kind of factual, comparison-friendly content AI engines cite confidently. Glossier and Rhode's content is more aesthetic and lifestyle-driven, which performs brilliantly on visual social platforms but gives AI engines less specific, citable substance to recommend from.

Key Finding #2: Mattresses Win, Luggage Loses

Of the five categories we tested, mattresses and sleep products had by far the strongest AI visibility, while luggage and travel brands lagged behind every other category.

Mattresses likely score highest because the category has an unusually mature third-party review ecosystem — Wirecutter, Sleep Foundation, GoodHousekeeping, and dozens of dedicated mattress review sites have spent a decade publishing detailed, comparison-heavy content. That density of independent, structured comparison content is exactly what AI engines draw on most confidently.

Key Finding #3: Negative AI Sentiment Is Its Own Category of Problem

Most low scores in our dataset reflect a brand simply not being mentioned. One brand's score revealed something different and more concerning: an AI engine actively surfacing negative sentiment, while the other two stayed neutral or positive on the identical query.

This is a distinct and arguably more urgent problem than simple invisibility. A brand that AI doesn't mention loses consideration passively. A brand that AI actively criticizes loses consideration and trust — and the buyer walks away having heard a specific, negative claim rather than no claim at all.

Key Finding #4: A Generic Brand Name Can Sink Your AI Visibility

One luggage brand in our study — sharing its name with a common calendar month — scored 17, the second-lowest in the entire dataset. Looking at the underlying responses, the brand wasn't being criticized. It mostly wasn't being recognized at all.

On category queries like "best carry-on luggage brands," the brand scored 0 across all three platforms — AI engines simply didn't connect the common word to the company. The single query where it scored well was a direct, branded comparison against a named competitor — context that forced disambiguation. Without that context, the brand name itself was a liability.

The Brands With Total AI Consensus

27 of the 50 brands scored within 10 points across all three platforms — meaning ChatGPT, Claude, and Gemini essentially agree on how visible (or invisible) they are. The standouts at the top of that list:

• Saatva — 98 (GPT 100 / Claude 95 / Gemini 100)
• Nectar — 97 (GPT 90 / Claude 100 / Gemini 100)
• Ilia — 97 (GPT 100 / Claude 90 / Gemini 100)
• Spanx — 93 (GPT 95 / Claude 90 / Gemini 95)
• Skims — 93 (GPT 95 / Claude 95 / Gemini 90)

Notably, Gemini scored brands highest on average across the dataset (74), ahead of ChatGPT (70) and Claude (68) — the reverse order from our earlier tech brands study, where Claude was the most generous platform. This suggests platform "bias" isn't a fixed personality trait — it varies by category, likely reflecting which platform's training data has denser coverage of a given industry.

Full Rankings: All 50 Brands

What This Means If You Run a DTC Brand

The biggest practical takeaway from this study: your social content strategy and your AI visibility strategy need to be planned separately, because they reward different things.

• Structured, factual content wins. The brands that scored highest — Saatva, Ilia, The Ordinary, Veja — all have extensive third-party coverage that explains ingredients, materials, or construction in comparison-friendly detail. Pure brand/lifestyle content doesn't give AI engines enough to cite confidently.
• Watch for negative sentiment, not just absence. A low score can mean two very different things — being ignored, or being actively warned against. If your brand has had a real quality or service issue, expect AI engines (particularly ones with more recent training data or retrieval) to surface it specifically when asked direct buyer-intent questions.
• If your name is a common word, work harder on association. Generic brand names need deliberate, repeated content pairing the name with the category — AI engines need that signal to disambiguate.
• Check all three platforms separately. Just as in our tech brands study, category-level "bias" shifted: Gemini was the most generous platform for DTC brands, while Claude was most generous for tech brands. There's no universal "best" platform to optimize for — you need your own data.