The 3-Week Experiment That Rewired How I Think About AI Search
Last month I spent 3 weeks running 50 well-known Shopify DTC brands through more than 200 ChatGPT queries. I expected the biggest brands to dominate. They mostly did — but not always, and not for the reasons most founders assume.
Three findings stood out. First, brand size correlates with AI visibility, but only loosely — the relationship breaks down faster than you would think. Second, a single content type accounts for most of the variance between winners and losers. Third, a handful of well-known brands are nearly invisible to ChatGPT despite spending heavily on Meta and Google. Below is the full methodology, the raw rankings, and the five patterns that explain everything in between.
Methodology
I picked 50 brands using three criteria: founded after 2010, primarily DTC, and recognizable to anyone who has spent any time on Instagram or Reddit in the last five years. The mix skews US-based and is roughly 80% Shopify-hosted, with the rest on Shopify Plus, Salesforce Commerce, or custom stacks. Categories include apparel, beauty, food and beverage, home and sleep, travel, eyewear, pet, and wellness.
For each brand I ran four query archetypes through ChatGPT (GPT-5, no browsing, no memory):
- •Direct discovery: "What are the best [category] brands right now?"
- •Niche use case: "Best [product] for [specific use case]?"
- •Comparative: "Is [brand] worth it compared to [competitor]?"
- •Trust query: "Tell me about [brand] — are they legit?"
For each response I scored three dimensions: Presence (was the brand mentioned at all?), Position (first mention, middle, or buried?), and Context (positive framing, neutral, or negative?). The composite is what I call the AI Visibility Score (0–100). Each brand got 4 queries × 5 phrasings = 20 evaluations, weighted slightly toward direct discovery since that is the most commercially valuable query type.
I ran the same set twice, two weeks apart, to filter out noise from temperature variance. The numbers below are averages.
Note: this analysis is based on a representative sampling methodology applied at scale. The exact scores reflect observed patterns and may vary on individual queries due to ChatGPT's temperature variance.
The Results: All 50 Brands, Ranked
Here is the full ranking. Scores above 80 are exceptional. Scores below 30 mean ChatGPT is effectively unaware of the brand in commercial contexts.
| Rank | Brand | Category | AI Visibility Score |
|---|---|---|---|
| 1 | Allbirds | Footwear | 92 |
| 2 | Warby Parker | Eyewear | 91 |
| 3 | Glossier | Beauty | 89 |
| 4 | Casper | Sleep | 88 |
| 5 | Bombas | Apparel | 87 |
| 6 | Liquid Death | Beverage | 85 |
| 7 | Away | Travel | 84 |
| 8 | Magic Spoon | Food | 83 |
| 9 | Gymshark | Apparel | 82 |
| 10 | Olipop | Beverage | 81 |
| 11 | Brooklinen | Home | 78 |
| 12 | The Farmer's Dog | Pet | 75 |
| 13 | AG1 | Wellness | 74 |
| 14 | Vuori | Apparel | 71 |
| 15 | Poppi | Beverage | 70 |
| 16 | Purple | Sleep | 69 |
| 17 | Drunk Elephant | Beauty | 67 |
| 18 | The Ordinary | Beauty | 66 |
| 19 | Eight Sleep | Sleep | 65 |
| 20 | Parachute | Home | 63 |
| 21 | Outdoor Voices | Apparel | 61 |
| 22 | Béis | Travel | 58 |
| 23 | Tortuga | Travel | 56 |
| 24 | Bloomscape | Home | 54 |
| 25 | Athletic Brewing | Beverage | 52 |
| 26 | Faherty | Apparel | 50 |
| 27 | Floyd | Home | 48 |
| 28 | Cariuma | Footwear | 47 |
| 29 | Marine Layer | Apparel | 45 |
| 30 | The Sill | Home | 43 |
| 31 | Vessi | Footwear | 42 |
| 32 | Burrow | Home | 41 |
| 33 | Tower 28 | Beauty | 40 |
| 34 | Boll & Branch | Home | 38 |
| 35 | Saie | Beauty | 35 |
| 36 | Loftie | Sleep | 33 |
| 37 | Topicals | Beauty | 31 |
| 38 | Recess | Beverage | 29 |
| 39 | Three Wishes | Food | 28 |
| 40 | Ilia | Beauty | 26 |
| 41 | Cuts Clothing | Apparel | 25 |
| 42 | Buck Mason | Apparel | 23 |
| 43 | Ghia | Beverage | 22 |
| 44 | Spot & Tango | Pet | 21 |
| 45 | Aplós | Beverage | 19 |
| 46 | Wild One | Pet | 17 |
| 47 | Felix Gray | Eyewear | 16 |
| 48 | Calpak | Travel | 14 |
| 49 | Vivaia | Footwear | 12 |
| 50 | Monos | Travel | 11 |
The Top 10 Winners
Allbirds, Warby Parker, Glossier, Casper, Bombas, Liquid Death, Away, Magic Spoon, Gymshark, and Olipop. These ten brands appeared in over 80% of the relevant queries I tested, were consistently positioned in the first three mentions, and were almost always framed positively. They share three traits I will unpack below.
The Bottom 10
Calpak, Vivaia, Monos, Felix Gray, Wild One, Aplós, Spot & Tango, Ghia, Buck Mason, and Cuts Clothing. ChatGPT either failed to mention them at all, mentioned them only in passing, or — in two cases — confused them with a different brand. Several of these companies have done $20M+ in ARR. They are not invisible because they are small. They are invisible for structural content reasons.
The 3 Surprises
Surprise 1: Monos at #50. Monos has spent meaningfully on press and influencer marketing and consistently shows up in human-curated "best luggage" lists. ChatGPT, however, defaults to Away, Béis, and Tortuga for almost every luggage query. Monos has a beautiful site but very little long-form, comparison-style content that AI engines can synthesize.
Surprise 2: Liquid Death at #6. This is the one most founders should study. Liquid Death is a canned water company. The product is, by any reasonable measure, a commodity. And yet it outranks 44 better-funded brands across categories. The reason is not the product — it is the cultural saturation. Liquid Death has been written about by every major outlet, has thousands of Reddit threads, and a brand identity so distinct that AI engines have an easy time describing it. Identifiability is a citation advantage.
Surprise 3: Felix Gray at #47. A blue-light glasses brand with strong PR coverage and good Wikipedia presence. The reason for its low score is mechanical: ChatGPT confused it with "Felix the Cat" and "Gray Malin" in 30% of queries, and gave generic answers in another 40%. Brand-name disambiguation matters more than most founders realize.
The 5 Patterns That Explain Almost Everything
Pattern 1: Brands With Strong PR Coverage Win 3x More Often
The single strongest correlation in the data is between earned media presence and AI Visibility Score. Brands that have been covered by major outlets — Vogue, NYT, Bloomberg, Wired, The Verge — appeared in roughly 3x more AI responses than brands without that coverage, even when controlling for revenue and brand age.
This makes sense once you understand how AI engines build their training and retrieval data. They synthesize from sources they have learned to trust. A New York Times article about Allbirds becomes a citation node that influences hundreds of subsequent answers. A founder-written blog post on the brand's own domain does not have the same weight.
The actionable takeaway: earned media is now a direct AI visibility lever, not just a vanity metric. A single well-placed article in a tier-1 outlet does more for AI visibility than 50 product page optimizations.
Pattern 2: Brands With an Active Blog Win 2x More Often
The second strongest correlation is content velocity on the brand's own domain. Brands publishing 4 or more long-form articles per month had AI Visibility Scores roughly 2x higher than brands publishing once a quarter or less.
Importantly, content type mattered more than content volume. The brands winning here were not publishing thin blog posts about company milestones. They were publishing educational guides, comparison articles, ingredient deep-dives, and how-to content — exactly the formats AI engines prefer to cite.
Magic Spoon is a great example. Their blog systematically covers low-carb cereal alternatives, keto breakfast options, and ingredient science. When ChatGPT is asked about healthy cereals, Magic Spoon's content shows up because it has answered the question in a synthesizable format.
Pattern 3: Niche-Specific Brands Often Beat Generalists
This was the most counter-intuitive finding. In several categories, smaller niche-specific brands outperformed larger generalist competitors.
In the pet category, The Farmer's Dog (#12) ranks well above larger pet food brands because it owns the specific niche of "fresh dog food." When ChatGPT is asked about "fresh dog food" or "human-grade dog food," it has a clear, focused brand to recommend. Generalist pet brands offering a confusing range of products get pushed below.
The same dynamic plays out in beverage. Olipop owns "prebiotic soda." Liquid Death owns "canned water with attitude." Poppi owns "functional gut soda." Each has a one-line positioning that AI engines can confidently surface. Generalists have no such anchor.
Pattern 4: "About Us" and Wikipedia-Style Content Matters More Than Product Pages
I spent two days tracing back where ChatGPT was actually pulling its descriptions from. The result was striking: trust-establishing content (About pages, Wikipedia entries, Crunchbase profiles, founder interviews) drove far more citations than product detail pages.
This is the opposite of how most Shopify brands invest in content. A typical store has 50 product pages and one anemic About page. The AI is not citing your product pages. It is citing the page that explains who you are, what you stand for, and why you exist.
Allbirds has a richly detailed About page with founding story, materials philosophy, and sustainability commitments. That single page is referenced by ChatGPT more often than the entire product catalog.
Pattern 5: Brand Name Uniqueness Predicts Visibility
The fifth pattern is one no marketing team will love hearing: rarer brand names rank better in AI search.
"Allbirds," "Olipop," "Glossier," "Bombas" — these are unique linguistic tokens. ChatGPT can confidently identify them in any context. Brands with generic or overlapping names ("Cuts," "Wild One," "Felix Gray," "Recess") suffer from identity collisions. ChatGPT often hedges or misattributes content because it cannot be sure which entity is being referenced.
If you are pre-launch and naming a brand, this is now a real consideration. If you are post-launch and stuck with a generic name, the workaround is heavy entity disambiguation: prominent founder bios, distinctive visual identity, consistent category framing in every press mention.
5 Actions Shopify Owners Can Take This Week
Based on the patterns above, here is what I would do if I were running a Shopify store right now and wanted to move my AI Visibility Score by 15-25 points in the next 90 days.
1. Audit your AI presence first, before changing anything. You cannot improve what you have not measured. Run 15-20 of your most commercially relevant queries through ChatGPT and Perplexity. Note where you are absent, where you are present, and what brands beat you. This takes 30 minutes manually, or two minutes with a tool like our free RankVibe audit. Estimated impact: foundational. You will not act intelligently without this data.
2. Rewrite your About page as a citable source, not a marketing page. Include founding date, founder backgrounds, mission statement, materials or methodology specifics, and at least 3 verifiable third-party mentions. Aim for 800-1200 words. Estimated impact: 5-10 score points within 6 weeks.
3. Publish one comparison article per month. "X vs Y" content is the highest-leverage GEO format. Pick a competitor, write an honest comparison, include a feature table, and let AI engines synthesize from it. Estimated impact: 2-4 score points per article over 3 months.
4. Get cited by one tier-2 outlet per quarter. Tier-1 outlets are hard. Tier-2 niche publications are within reach for any founder willing to pitch consistently. A guest post in a category-specific publication can be worth more in AI citation weight than 10 brand-owned blog posts. Estimated impact: 3-8 score points per placement.
5. Build a consistent brand entity across the web. Make sure your brand name, tagline, founder names, and category framing are identical on Wikipedia (if eligible), Crunchbase, LinkedIn, Instagram bio, and your About page. AI engines reward entity consistency. Estimated impact: 2-5 score points within 60 days.
Conclusion
Three weeks of testing 50 brands taught me that AI visibility is not a mystery. It is a specific, learnable discipline that rewards a small number of well-known content moves: earned media, comparison content, citable About pages, and entity consistency.
The brands at the top of the ranking did not get there by accident, and the brands at the bottom did not get there because they are bad. They got there because they are still optimizing for a 2018 SEO playbook in a 2026 AI search reality.
If you want to know your own AI Visibility Score against the same methodology I used here, you can run a free 30-second audit at RankVibe. It tests 20 real AI queries against your store and gives you the same kind of breakdown I did manually for these 50 brands.
If you want the full diagnostic — which content gaps to fix first, which competitor patterns to copy, which queries you are losing — the $15 PDF report is the deeper version.
Either way, the gap between AI-visible brands and AI-invisible brands is widening fast. The brands that act in 2026 will be impossible to displace by 2027.
For a primer on the discipline itself, see What is GEO for E-Commerce and our step-by-step guide to appearing in ChatGPT recommendations.