Channel Disagreement Playbook: Which Feed Fields to Sacrifice When AI Shopping Platforms Conflict
The conflict is real. ChatGPT Shopping, Perplexity, and Google AI Overviews each pull different signals from your product feed, and when they disagree, you can't win all three at once.
I've seen this firsthand across 40+ Shopify store audits. You tune your titles for Google's keyword requirements, and they read like garbage to ChatGPT's natural language engine. You write rich descriptions for AI recommendations, and you lose the keyword density that Google still rewards. Something has to give.
This is the playbook for deciding what to sacrifice, and how to make that trade-off based on your actual business goals, not platform hype.
Why Do AI Shopping Platforms Disagree on Feed Fields?
Each platform has a different model of what a product is. That's the root of every conflict you'll hit in your feed.
Google Merchant Center was built around paid search. Its product data specification lists 50+ required and optional attributes, many of them holdovers from an era when keyword match determined relevance. Titles up to 150 characters, keyword-heavy descriptions, taxonomy aligned to Google's own product category tree. The system rewards density.
ChatGPT Shopping runs on OpenAI's commerce feed spec, which introduces fields like enable_search and enable_checkout, signals about intent and purchase flow. Title and description requirements lean toward natural language. OpenAI's system is trying to match conversational queries like "find me a quiet cordless vacuum for a small apartment" to products. Verbose keyword strings actively hurt that match quality.
Perplexity is a web-crawler-first system. It doesn't publish a formal feed spec at all. It surfaces products by reading both structured feeds and product pages directly, rewarding review mentions, spec clarity, brand authority signals, and how well your product pages answer questions. On-page schema matters here as much as feed data.
Three systems. Three different ideas of what "good product data" looks like. You're stuck in the middle.
What Are the Most Common Feed Field Conflicts?
Four fields cause most of the headaches. Everything else is secondary.
1. Product Title
Google wants keyword density. Their spec allows 150 characters and actively recommends including material, color, size, and model number. "Women's Running Shoes Nike Air Zoom Pegasus 41 Blue Size 8" is a Google-optimized title.
ChatGPT performs better with intent-matched, natural language titles. That same shoe might serve better as "Nike Air Zoom Pegasus 41, Women's Cushioned Running Shoe." Shorter, cleaner, easier for the model to match to conversational queries.
You can't write one title that wins both.
2. Product Description
Google's algorithm still rewards keyword repetition in descriptions. Five hundred to 1,000 words of dense product copy can improve Shopping ad relevance scores.
ChatGPT and Perplexity want descriptions that answer questions. Think: "Who is this for? What problem does it solve? What makes it different?" A 150-word description that answers those three questions outperforms a 600-word keyword block for AI recommendation engines. Every time.
3. Category Taxonomy
Google has its own Product Taxonomy, a 6,000+ item classification tree you're expected to match your products against. ChatGPT Shopping uses its own category schema. Perplexity uses neither, inferring category from semantic context instead.
If you're classifying products for Google's taxonomy, you may be using categories that mean nothing to OpenAI's commerce system. And vice versa.
4. Custom Attributes and Supplemental Feeds
Google supports supplemental feed data for custom labels (used in bidding segmentation), sale prices, and promotion IDs. OpenAI's feed spec supports different supplemental fields: enable_search, enable_checkout, and variant-level attributes.
These don't overlap. Populating one set of custom fields doesn't help the other platform at all.
How Do You Decide Which Platform to focus on?
Here's the framework. Start with where your customers actually are right now.
If 80% of your traffic and revenue still comes from Google Shopping, don't blow up your feed optimization for ChatGPT. Not yet. ChatGPT Shopping's user base is growing fast, but Google still moves more product for most Shopify stores in 2026. Chasing the future platform at the cost of your current revenue is a bad trade.
Once you've grounded the decision in current data, layer in margin and brand goals:
| Product Type | Primary Platform | Why |
|---|---|---|
| High-margin, brand-driven | ChatGPT Shopping or Perplexity | Recommendation flows reward brand narrative and quality signals over price. Being cited by ChatGPT to a buyer asking "what's the best camera bag for travel photographers" is worth more than a Google Shopping impression. |
| Commodity or price-competitive | Google AI Overviews / Shopping | High-volume, price-sensitive categories still convert best through Google's visibility and price-match surfaces. A $24 kitchen gadget sells on price and reach, not story. |
| New product launches | Perplexity | Perplexity's web-crawler model picks up new product pages faster than Google indexes and ranks them. A well-structured product page with solid schema can surface in Perplexity within days of launch. |
What Does the Field Priority Map Actually Look Like?
Stop trying to win everything. Pick a primary platform per product segment and go deep on it.
If Your Primary Is Google AI Overviews / Google Shopping
- Title: 150 chars, keyword-first, include attributes (color, size, material, model number)
- Description: 500-1,000 words, keyword-rich, technical specs included
- Category: Match Google's Product Taxonomy precisely
- Required fields: GTIN, MPN, brand, condition, availability, price
- Worth adding: Shipping info, return policy, product ratings feed
If Your Primary Is ChatGPT Shopping
- Title: 80-100 chars, natural language, intent-matched to how buyers ask questions
- Description: 100-200 words answering "who is this for, what does it do, what makes it different"
- Category: Match OpenAI's commerce schema, not Google's taxonomy
- Required fields:
enable_search: true,enable_checkout: true(if applicable), clean image URL, price, availability - Worth adding: Review summary text, variant-level data, brand description field
If Your Primary Is Perplexity
- Title: Descriptive and specific, Perplexity matches on semantic meaning, so be precise
- Description: Question-answering format. Lead with the use case, then specs, then differentiation.
- On-page priority: Product page schema markup (Product, Offer, Review) matters as much as feed data. Perplexity reads both.
- Required: No formal spec to match, so focus on web-crawlable structured data and on-page Product schema
- Worth adding: Third-party review mentions, spec comparison tables, direct citations from authoritative sources
According to Shopify's 2026 guide on optimizing for Perplexity Shopping, product identifiers like GTINs, MPNs, and Google product categories are also factors Perplexity uses to resolve product identity across the web. Those fields pull double duty, run them for Google, get partial credit on Perplexity.
Can You Build a Feed That Partially Satisfies All Three?
Sort of.
These six fields are non-negotiable regardless of primary platform:
- A clean, accurate product title (even if length and style vary by platform)
- At least one high-resolution image (1000x1000 or larger)
- Real-time accurate price and availability
- GTIN where you have it
- Brand field populated, not blank
- A product description you wrote yourself, not copy-pasted from the manufacturer
Everything above this baseline is where you make trade-offs. And one practical move: use a supplemental feed layer.
Your master feed gets the Google-optimized version of titles and descriptions. A supplemental feed or direct API integration pushes the ChatGPT-specific fields with natural language versions. Shopify's metafields work well for storing alternate content without breaking your primary feed structure.
It's extra work. Worth it if you're pushing volume through multiple AI channels and the margin is there to justify the maintenance overhead.
What Should You Actually Do This Week?
One move. Not five.
Pick your highest-margin product category. Pull those products' current feed data. Run them against the field priority map for your primary platform. Identify the three biggest gaps, usually title format, description length and format, and category taxonomy. Fix those three fields for that one category first.
Then measure. ChatGPT Shopping attribution shows up in referral traffic from openai.com. Perplexity referrals come in from perplexity.ai. Add UTM parameters to your product URLs so you can separate Google Shopping clicks from AI platform clicks in your analytics. Without that tracking in place, you're flying blind on whether any of this is working.
I've run this process on dozens of stores. The ones that see results fastest aren't doing the most work. They're doing the most focused work on the right platform for their product mix.
That's the whole game.
Is Your Store Ready for Agentic Commerce?
AI shopping assistants are already recommending products, and most Shopify stores aren't showing up. Find out where your product feed stands and what to fix first.
See the Agentic Commerce ChecklistFrequently Asked Questions
Do I need separate product feeds for ChatGPT Shopping, Perplexity, and Google?
Not necessarily separate feeds, but you may need supplemental feed layers. Your master feed can serve your primary platform (usually Google). For ChatGPT Shopping, OpenAI's commerce feed spec defines additional fields like enable_search and enable_checkout that don't exist in Google Merchant Center. Shopify metafields work well for storing alternate content without breaking your primary feed.
Which feed field matters most for ChatGPT Shopping recommendations?
Product title and description matter most, but style matters as much as content. ChatGPT's recommendation engine matches conversational queries to products, so natural-language titles (80-100 characters) and concise descriptions that answer "who is this for, what does it do, what makes it different" outperform keyword-dense copy. The enable_search field must be set to true or your product won't appear in ChatGPT Shopping at all. Full stop.
How does Perplexity Shopping differ from Google and ChatGPT in how it reads product data?
Perplexity is a web-crawler-first system with no formal published feed spec. It surfaces products by crawling product pages and reading on-page structured data (Product, Offer, and Review schema). Your product page quality matters as much as your feed data. Third-party review mentions and spec clarity also factor into Perplexity recommendations in ways that don't apply to Google Shopping at all.
What's the minimum viable product feed that works across all three AI platforms?
Six fields are non-negotiable across all three: a clean accurate product title, at least one high-resolution image (1000x1000 or larger), real-time accurate price and availability, GTIN where available, a populated brand field, and a product description you wrote yourself. Not the manufacturer's copy. Everything beyond this baseline is where platform-specific trade-offs begin.
Should I improve for ChatGPT Shopping if most of my revenue still comes from Google?
Don't blow up your existing Google feed optimization to chase ChatGPT traffic. For most Shopify stores in 2026, Google still moves more product. The practical approach is to keep your Google-optimized master feed intact and layer in ChatGPT-specific fields through supplemental data or Shopify metafields. Start with your highest-margin product segment, where ChatGPT recommendation traffic has the biggest revenue impact per conversion.
