A runner doesn’t ask one question. They ask ChatGPT for the best beginner shoe, the best for wide feet, the best on a budget, the best for marathon training — and each time they get a different shortlist. The headline from running across all of them: there is no single “best running shoe” brand in ChatGPT’s eyes. The winner changes completely with the intent.
Nike led budget and long-distance; ASICS led the generic “best running shoes”; Brooks led beginners; HOKA led wide feet. For a brand, that’s the whole game — you don’t need to win the category, you need to own an intent.
| Brand | Led | Avg share | Sentiment | Score |
|---|---|---|---|---|
| Nike | 5 | 32% | +1.4 | |
| ASICS | 3 | 26% | +1.3 | |
| Brooks | 2 | 22% | +0.8 | |
| New Balance | 1 | 23% | +0.9 | |
| HOKA | 1 | 23% | +1.2 | |
| adidas | 0 | 22% | +1.3 |
Five questions, five different winners
Run the category the way a runner shops and the leaderboard refuses to settle. The generic “best running shoes” went to ASICS (it led two of three runs, Nike the third). “Best running shoes for beginners” went to Brooks (two of three). “Best budget running shoes” went to Nike (three of four, New Balance once). “Long distance” went to Nike, “wide feet” to HOKA. None of them is the obvious mega-brand for that need — each won because the sources framed it as the answer to that specific question.
“You don’t need to win running shoes. You need to win one of the questions runners actually ask.”
It judges you on the runner’s goal, not your marketing
Look at what ChatGPT rewarded for beginners: cushioning and joint protection — not getting hurt — not the speed and performance shoe brands usually lead with. The deciding quality changes with the question, and it’s frequently one you’d never put on a billboard. Win that quality in the reviews, even an unglamorous one, and you’re a candidate; miss it and you’re invisible however good the shoe is.
Runner’s World and Reddit decide it
The answers were built from running-specialist publications and Reddit — Runner’s World, RunRepeat, Running Gear Lab, RTINGS — not brand sites, and barely any mainstream outlets (Forbes appeared once). Reddit was the second most-cited source overall, carrying as much weight as the magazines. Getting into those round-ups and threads is the highest-leverage move; your own website barely features. In AI search, earning citations in the sources the model reads is what generative engine optimization (GEO) actually looks like for an apparel brand.
ChatGPT rewrites the question before it searches
It rarely searches the words typed. It adds the year and “reviews,” and on “best running shoes for long distance” it reframed the search as “marathon training” entirely — injecting the deciding context itself. You’re optimising for ChatGPT’s rewrite, not the runner’s phrase, and that rewrite leans toward review language and specialist terms.
“On ‘long distance,’ ChatGPT didn’t search your words — it searched ‘marathon training.’”
The buy box is real — and Amazon isn’t on it
Shoes trigger a full shopping panel, and the sellers split by intent: premium picks come from Dick’s, Zappos, Fleet Feet and Scheels; budget picks from DSW, Kohl’s, Famous Footwear and Big 5. Across every run, the three biggest US retailers — Amazon, Walmart and Target — never appeared. Your buy-box visibility depends on clean feeds at this specific, non-obvious set of partners.
And the price you see isn’t yours. The same Nike Alphafly 3 ranged $176.97 to $304.99 in a single answer — a 72% spread — because ChatGPT surfaces whichever integrated retailer’s offer ranks highest on price, availability and seller status. Whoever has the cleanest, best-priced feed at the right retailer wins the in-chat sale.
Ask again, and the leader can move
Repeating the headline intents shows the order isn’t fixed: “budget” went to Nike three of four runs (New Balance once); “best running shoes” split ASICS two, Nike one; “beginners” went Brooks two, ASICS one. Strong patterns, but a single check would mislead — visibility has to be sampled, not snapshotted.
What to do if you sell an apparel product
Own an intent, not the category
No brand wins everything. Pick the intents you can credibly win — your beginner shoe, your wide-fit, your value line — and concentrate there. That’s how Brooks owns beginners and HOKA owns wide feet.
Win Runner’s World, RunRepeat and Reddit
Those are the sources ChatGPT reads. Earned placements and genuine community presence move the answer; your own site barely features. Treat it like PR, not on-site SEO.
Aim to be championed, not just listed
Nobody is criticised in shoes, so the edge is warmth of description. Reviews that call you “the best for X” beat being one more name in a list.
Get feeds into the right retailers
Not Amazon — the shelf is Dick’s, Zappos, Fleet Feet and Scheels for premium, DSW, Kohl’s and Famous Footwear for budget. Make sure your best-priced stocked retailer has a clean, complete feed.
See what AI says about your brand
Run your own prompt through the live Chaice console and get the same breakdown — attributes, brands and sources — in seconds.
Try the console