Every platform arranges product information differently — Google wants particular column names, Meta wants different field names and different image rules, and OpenAI's commerce schema for ChatGPT Ads shares almost no field names with either. Format conversion, in Advanced feed tools, takes a feed built for one platform and reshapes it for another, so putting your existing Google feed on ChatGPT — or moving it to Meta — does not require rebuilding it from scratch. Columns come from the destination's own published specification rather than a renamed copy of the source's, and before you convert anything it tells you which columns carry across, which get rewritten, which are derived, and which nothing can supply — naming the last group instead of shipping them empty. When the destination is a feed-file-only channel, it says so up front.
Adding a channel usually means rebuilding your feed. The product selection is the same, the catalogue is the same, but the destination wants different column names, a different price format and a different category system — so the work starts again from a blank sheet.
The dishonest version of this tool is worse than no tool. Rename the columns you can, leave the rest empty, report the migration as complete — and the merchant finds out when the destination rejects the feed, days later, with no indication of which column was the problem.
How it works
Choose a source and a destination
Google Merchant Center, Meta Catalog and ChatGPT Ads, in any direction. A platform appears only when its mapping is complete against that platform’s own published specification — otherwise you could "migrate" onto a feed missing twenty required columns, which is the failure this exists to prevent.
Columns come from the destination’s spec
The plan is derived from what the destination will actually emit, not from a renamed copy of the source. That is why a Google → ChatGPT conversion produces item_id and image_url rather than Google’s field names, and why the ChatGPT category is a free-form path rather than a Google taxonomy value.
Every column is classified before you run it
Carried means it transfers unchanged. Rewritten means the destination words the same thing differently, and the value is translated. Derived means it is computed. Unavailable means nothing in your data can supply it — and those are named, with what to do instead, rather than shipped as empty columns.
The analysis counts feed rows, not products
One product in five sizes is five feed rows, so ready and blocked are counted over rows. A problem affecting every row of one product is reported as one product’s worth of work rather than five, and per-column fill coverage makes an empty column visible instead of silent.
Export, then connect
Export as CSV, XML or JSON. Google and Meta can be pushed from Push to channels behind the usual per-change approval. A feed-file-only destination such as ChatGPT Ads gives you the validated file and a hosted URL, and says so before you start so there is no assumption to correct later.
What you get
- ✓Conversion between Google Merchant Center, Meta Catalog and ChatGPT Ads, in any direction
- ✓A column-by-column plan — carried, rewritten, derived, unavailable — before anything runs
- ✓Unavailable columns named individually, with the action that resolves each one
- ✓Row-level ready and blocked counts, with one problem grouped across many products
- ✓Per-column fill coverage, so an empty column cannot pass unnoticed
- ✓Filters, sort and limit carried across from an existing feed configuration
- ✓A file-only warning when the destination is one FeedGraph does not deliver to
- —It converts between Google, Meta and ChatGPT Ads only. Amazon, Shopify and custom output are excluded — the first two because their channel is not open, custom because it has no destination spec to validate a conversion against.
- —It does not read your existing feed file. The data comes from your FeedGraph catalogue; choosing a source platform tells FeedGraph which vocabulary you are coming from, so the plan can explain what changes.
- —Some columns genuinely cannot be produced. Meta’s own product taxonomy has no published mapping from Google’s, so the feed carries google_product_category as the fallback Meta accepts and tells you to set Meta’s field in Commerce Manager if you want its own categorisation. Naming that is the point — the previous implementation wrote it blank and called the migration complete.
- —A completed conversion is not a live channel. It produces a validated feed; delivery is Push to channels for Google and Meta, and a hosted URL you connect yourself for ChatGPT Ads.
Questions
Can I turn my Google feed into a ChatGPT Ads feed?
Yes, and that is one of the three supported directions. OpenAI’s commerce schema shares almost no field names with Google’s, so the conversion is a genuine translation rather than a rename — and before you run it you see which of your columns carry across, which get rewritten, and which nothing can supply. seller_name is the one field no other schema has an equivalent for; you set it once as an account setting.
What happens to a column the destination needs but my data cannot fill?
It is listed as unavailable, by name, with the reason and what to do instead. It is never written as an empty column, and the conversion is never reported as complete while a required column is unfillable.
Will my product selection and filters carry across?
Yes. Point the conversion at an existing feed configuration and its filters, sort order and limit come with it, so you are converting the same set of products rather than rebuilding the selection.
See Format conversion in use
More in Distribution
All features →Channel feeds
Create and manage product feeds for Google, Meta and ChatGPT Ads — plus custom output in any format. Amazon and TikTok coming next.
View feature →Advanced feed tools
The guided builder: choose products, map fields to platform attributes, apply transformations, and produce a finished file.
View feature →Segments
Product groups that maintain themselves — by rule, or hand-picked — usable in any feed, campaign or report, and publishable as a live link.
View feature →Ready to try it?
FeedGraph is in early access. Apply to join and we'll onboard you personally — no credit card required.
Join early access