The list of products you send to a platform like Google or Facebook. It contains all the details each platform needs — titles, prices, images, stock status — formatted the way that platform expects.
Your one main, up-to-date list of products inside FeedGraph. Everything else — every feed you send, every ad you run — is built from this single source. Fix something here and it's fixed everywhere.
A place you sell or advertise — Google, Meta (Facebook and Instagram), ChatGPT Ads, your own website, and marketplaces like Amazon or TikTok Shop. Each channel has its own rules about how product data should be formatted. FeedGraph publishes to Google and Meta today, plus custom output in any format. For ChatGPT Ads it builds, validates and hosts the feed file and you connect it in OpenAI's Ads Manager. Amazon and TikTok are coming next.
A product that has everything filled in correctly and will be accepted by a platform without being rejected. FeedGraph marks products as feed-ready when they pass all required validation checks for a given channel.
A serious problem that will stop a product going live — for example, a missing price or no product image. Blocking issues must be fixed before a product can be included in an export. FeedGraph shows them in red.
A smaller issue that won't block a product from going live but could hurt how well it performs — for example, a title that's too short, or a missing GTIN. Warnings are shown in amber.
A single piece of product information — title, price, colour, size, image URL, brand, and so on. Different platforms require different attributes, and some attributes are mandatory while others are optional.
A saved group of products defined by a rule — for example, "best sellers" or "high-profit items" or "all apparel currently in stock". A segment updates itself automatically as products come and go, so it's always current.
A hand-picked set of products built for a specific campaign — for example, "Summer Sale" or "Seasonal gifting under $50". Where a segment is rule-based and automatic, a cluster is curated and stays fixed until you change it.
A word or short phrase you attach to products so you can group them inside Google or Facebook ads. Labels let you apply different bidding strategies to different product groups — for example, a higher bid for "bestsellers" than for "clearance".
Keeping FeedGraph and your connected store (or ad platforms) showing exactly the same, up-to-date information. When everything is in sync, the price on your website, in FeedGraph, and in your Google feed all agree.
How much revenue you make back for every $1 you spend on ads. A ROAS of 4 means you make $4 in sales for every $1 spent. FeedGraph shows ROAS per product so you can see exactly which SKUs are earning and which are wasting budget.
A product a platform has turned down, so it can't appear in ads until the issue is fixed. Google calls this "disapproved". Meta calls it "rejected". FeedGraph's GMC Recovery Center collects all disapproved products in one place and explains why each one was rejected.
The unit of account for AI features in FeedGraph. Things like AI writing, image generation, and AI assistant queries use credits. 1 credit = $0.10. Subscription plans include a monthly allowance; additional credits can be purchased as top-ups that never expire.
A channel FeedGraph builds and hosts a validated feed for, but does not deliver to and reads no results back from. ChatGPT Ads is the one today: you take the hosted feed URL and connect it in OpenAI's Ads Manager. It is deliberately not described as "live" alongside Google and Meta, because FeedGraph neither pushes to it nor reports on its spend.
OpenAI's advertising product inside ChatGPT. Its product-feed campaigns run from a catalogue feed written in OpenAI's own commerce schema — its own field names, a price written as a single string like "79.99 USD", and a free-form category path rather than the Google taxonomy. It has no keywords, so ad-group context hints carry the relevance instead. Products expire 14 days after they drop out of a feed snapshot, which is why a hosted feed URL rather than a one-off upload is the only reliable way to run it.
Short phrases attached to a ChatGPT Ads ad group that tell it which conversations your products are relevant to. They are the equivalent of keywords on a channel that has none. FeedGraph drafts them from the products actually in your selection rather than in general terms.
Reshaping a feed you already maintain into another platform's format — Google Merchant Center, Meta Catalog or ChatGPT Ads. Columns come from the destination's own published specification, and each one is classified as carried across, rewritten into the destination's wording, derived, or unavailable. Unavailable columns are named individually rather than exported blank.
A score FeedGraph gives each product showing how ready it is for a specific platform — taking into account required fields, recommended fields, image quality, and platform-specific rules. A higher score means fewer likely rejections and better ad performance.