How do I know if my products are in the Shopify Global Catalog?

There is no toggle in your admin that confirms it. The reliable way is to query the catalog the way assistants do, using a specific product identifier or a tight description, and see whether your item comes back with the right title, price, and image. If it returns under a Universal Product ID with your offer attached, it has been absorbed. If a category search surfaces competitors but never you, something is stranded.

How long does it take for a new product to appear in the Global Catalog?

It is not instant and Shopify does not publish a fixed SLA. Ingestion runs on Shopify's schedule, so a new or edited product can take from hours to several days to settle into the catalog. That lag is exactly why you should record a baseline before a change and re-check on a schedule rather than expecting same-day results.

Why would a product be missing from the catalog even though it's published?

Common causes are an unmapped or wrong product category, missing identifiers like GTIN, thin or empty key fields, an out-of-stock or unbuyable variant, or markets and channel settings that exclude the product from agentic surfaces. Published on your storefront and present in the cross-merchant catalog are two different states.

Can I prove that fixing my catalog data actually lifted my AI visibility?

Yes, if you measure deliberately. Fix one variable, keep a fixed list of queries, and record presence, rank, and which competitor beats you before and after, with dates. The trend across repeated checks is the proof. A single snapshot is noise; a logged series against the same queries is evidence.

Did the Shopify Global Catalog Absorb Your Products? How to Check and Prove the Lift

Publishing a product to your Shopify storefront and getting that product into the Global Catalog are two different events, and merchants keep confusing them. Your product page can be live, indexed by Google, and selling, while an AI assistant asked for exactly that item still recommends three competitors and never mentions you. The missing step is ingestion: the catalog has to absorb your product before any assistant can surface it.

There's no banner in your admin that says "absorbed." So you have to check it yourself. Here's how to confirm whether the Global Catalog has your products, find what's stranded, and build a record that proves the lift when you fix things.

Why "published" and "in the catalog" are not the same

Shopify's Global Catalog is a cross-merchant index that ChatGPT, Gemini, and Shop search over when a shopper asks for a product. It is not your storefront and it is not Google's index. It is a separate layer that ingests product data from across Shopify, clusters identical items under a Universal Product ID, and attaches each seller's offer underneath. The mechanics of how it clusters and chooses are in the agentic buy box, and the wider picture is in the Global Catalog guide.

Absorption is the moment your product becomes a record in that index with usable fields. Until then, an assistant has nothing to surface. A product can be perfectly live for human shoppers and effectively invisible to agentic ones, and nothing in your dashboard tells you which state you're in.

The absorption check, three passes

You don't need special access to test this. The catalog is queryable, which is how assistants read it, so you can run the same kind of lookups and read the results back.

Pass one: identity lookup. Take a product you know well and search for it precisely. Use a unique part of its title, or better, a strong identifier like a GTIN. If the catalog returns your product with the right title, price, and image, it has been absorbed. If it returns a similar product from someone else, or nothing, that's your first signal.

Pass two: category search. Now search the way a stranger would, by category rather than brand, like "merino base layer" or "stainless dog bowl." This tells you something the identity lookup can't: whether you're absorbed and discoverable, or absorbed but buried. Plenty of products are in the catalog and still never surface for the searches that bring new customers. If you want the full method for this, the rank-checking walkthrough in how to see where your store ranks covers it.

Pass three: clustering check. For any product other stores also sell, look at whether yours is grouped under the same identity as theirs or stranded on its own. A mis-clustered product is technically absorbed but cut off from the comparison that actually drives sales. This is where identifiers earn their keep, because clean GTIN and MPN data is what lets the catalog match your item to the right cluster.

A faster way to run all three at once is our free Catalog Readiness Index check, which tells you what the catalog can and can't read from your products before you start hand-querying.

What gets a product stranded

When a product fails the absorption check, the cause is almost always one of a short list:

Wrong or missing product category. If the catalog can't place your product in the taxonomy, it struggles to ingest and cluster it. We dug into why this single field matters so much in product taxonomy as a ranking factor.
Missing identifiers. No GTIN or MPN means no clean way to match your item to the right product cluster, so it strands or duplicates.
Thin fields. Empty or vague titles, descriptions, and attributes give the catalog little to index. The fields that actually move the needle are listed in the product fields that decide your rank.
Unbuyable state. Out of stock, not shipping to the shopper, or a variant that can't check out cleanly. An offer that isn't buyable is one the assistant routes around.
Channel and market exclusions. Settings that keep a product off agentic surfaces will keep it out of the catalog's usable set even when it's live on your theme.

The reason this matters: each cause has a different fix, and you can't pick the right fix until you know which one is hitting you. That's the whole point of checking rather than guessing.

Proving the lift over time

Confirming absorption once is useful. Proving that your fixes worked is what turns this into a discipline you can defend to a client or a boss.

The method is boring and that's why it works. Pick a fixed list of queries, maybe ten, mixing identity lookups and category searches. Run them, and for each one record three things: whether you appear, where you rank, and who beats you. Date the record. Then change one variable, like adding GTINs to a product line or rewriting a category-mapping, and wait long enough for the catalog to re-ingest. Ingestion runs on Shopify's schedule, so give it from a day to several before you re-check.

Run the same query list again, log it next to the first. Now you have a before and after on the same queries with dates, which is evidence rather than a vibe. Do it monthly and the trend line tells the real story, because catalog data drifts as your inventory, pricing, and competitors move whether you touch anything or not.

Two rules keep the data honest. Keep the queries identical every time, or the numbers aren't comparable. And change one thing at a time, or you'll never know which fix earned the lift.

Where this leaves you

The Global Catalog is the layer that decides whether AI assistants can see your products at all, and it operates silently. You don't get a notification when you're absorbed, when you're stranded, or when a competitor's better data quietly pushes you out of a cluster. The only way to know is to look, and the only way to prove progress is to look the same way repeatedly and write it down.

That's the loop we built AgentReady around: confirm what the catalog can read, fix what's stranded, and track the lift instead of hoping. If you want to see where you stand right now, start with the Catalog Readiness Index check, then see how the full Catalog tier closes the loop on winning the AI buy box.