How ChatGPT chooses brands for ecommerce answers

Shopify brands ask: “How do we get into ChatGPT’s answers?” There is no public ranking factor sheet. There is a repeatable pattern across audits.

The working model

ChatGPT-style answers usually need four conditions at once:

1. Entity clarity — the model knows who you are (consistent name, domain, Organization schema)
2. Fetchability — GPTBot / related agents can read public HTML (allow crawlers)
3. Product machine-readability — Product JSON-LD, offers, identifiers (Shopping requirements)
4. Citable proof — on-site facts plus third-party mentions (citability)

Miss one layer and you can still appear sometimes. Miss two and you become invisible on competitive prompts.

Discovery vs Shopping vs browsing

• Chat answers — shortlists and advice in conversation
• ChatGPT Shopping — product cards that lean on feed-like completeness
• On-behalf browsing — ChatGPT-User fetching a page the shopper asked to open

Optimize for all three; do not assume Shopping eligibility equals answer SoV. Definitions: ChatGPT Shopping, GPTBot.

What ChatGPT appears to prefer in product categories

• Direct answers (“best for X if you need Y”)
• Concrete attributes over vibe copy
• Brands that show up in multiple reputable sources
• PDPs that are not trapped behind quizzes or challenge walls

Shopify-specific failure modes

• Security apps challenging all bots
• Theme apps injecting broken/duplicate Product JSON-LD
• Brand name mismatch (legal entity vs storefront vs packaging)
• llms.txt on *.myshopify.com while customers use the custom domain

Playbook

1. Unblock AI crawlers and verify PDP HTML
2. Fix Organization + Product schema on hero SKUs
3. Publish answerable collection intros and FAQs
4. Ship llms.txt as orientation, not a magic file
5. Measure on a fixed panel (methodology)
6. Earn citations where your category researches

Full checklist: Get recommended by ChatGPT.