Connect a Product Catalog

Product-level shopping metrics start when a catalog is connected. Matching is retroactive: on connect, observations from the last 30 days are matched immediately, so the tables fill from data already collected instead of waiting for the next window.

Choose a connection

Shopify. Paste your store domain. The public products.json listing is read directly, so there is no app to install and no login to share. The catalog re-syncs nightly.

Google Merchant Center feed. Paste the URL of the product feed you already submit to Merchant Center (CSV, TSV, or JSON). The same feed you maintain for Google Shopping powers Citlyze unchanged, and it re-syncs nightly.

CSV upload. A minimal file works: title and brand columns are required; price, currency, link, image_link, category, gtin, and mpn improve matching and unlock price metrics. Uploads do not re-sync automatically; upload again when the catalog changes.

Re-syncs are additive: products missing from a new sync are deactivated, never deleted, so historical matches keep resolving. Plan limits cap the number of active catalog products; the SKU pack add-on raises the cap.

How matching works

Observations resolve to catalog products deterministically:

• Exact: a GTIN or MPN match, or the same brand with an identical normalized title.
• Strong: the same brand with heavily overlapping title terms.
• Fuzzy: high title similarity where the brand signal is weaker.

Anything below these thresholds stays unmatched rather than guessing, and unmatched observations still count in merchant and prompt metrics.

Make your feed work harder

The same fields that improve Merchant Center performance improve AI shopping visibility, because Google's shopping surfaces resolve products against Shopping data:

• Fill gtin (or mpn when no GTIN exists) on every product. Identifier matches are what tie an AI carousel entry to your SKU with certainty.
• Keep title specific: brand, product line, and variant beat generic names both for ranking and for matching.
• Provide image_link, price, and availability, and keep them current; stale prices show up directly in the merchant price comparison.
• Use full category paths in google_product_category.

Publish an OpenAI product feed

When ChatGPT capture is active, publishing OpenAI's product feed gives ChatGPT first-party product data instead of leaving your products to third-party sources. The feed is a push-based file (TSV, CSV, XML, or JSON) following OpenAI's commerce feed specification, refreshed on a short cadence, with required seller and product fields largely mirroring a Merchant Center feed. If you already maintain a Merchant Center feed, most of the work is done; export the same data in the OpenAI format and keep both fresh.