Clean Up Your Catalog

Why this matters

A product catalog imported from Shopify or a spreadsheet almost always has gaps: missing brand names, vague descriptions copied from a supplier, no product category, wrong apparel attributes. These gaps prevent products from being exported to platforms like Google and Meta, and they silently reduce ad performance on products that do make it through.

The Needs fixing queue is designed to fix all of this quickly — in bulk, with AI doing the heavy lifting.

Step 1: Open the Needs fixing queue

Open All products and switch to the "Needs fixing" tab. FeedGraph automatically scores your catalog and groups every problem by type — for example: "342 products missing a description", "88 products missing a Google Product Category", "24 products with invalid image URLs".

Step 2: Pick a problem group to tackle

Start with the most common or highest-impact issue — typically missing descriptions or missing product categories, since these affect feed eligibility for Google and Meta. Click the group to see the affected products.

Step 3: Fix with AI

Click "Fix with AI" for the selected group. FeedGraph generates the missing values for every product in the group simultaneously — drafting descriptions, assigning categories, or extracting attributes from existing product data.

A credit estimate is shown before the job runs.

Step 4: Review and approve

Enriched values appear alongside the original in a review interface. You can:

• ▸Approve all — accept everything in one click after spot-checking a few
• ▸Approve individually — go product by product if you want tighter control
• ▸Reject and manually edit — override any suggestion that doesn't fit your brand

All changes are logged with a full audit trail showing what was changed, when, and by what operation.

Step 5: Move to the next problem group

Repeat until your catalog is clean. FeedGraph updates the health summary in real time — you'll see the number of ready products increase as you work through each issue group.

Tips

- Run category assignment before description generation — correct category data improves the AI's description quality.

- Use the "filter to this brand" option to fix problems one brand at a time if you manage a multi-brand catalog.

- After your first clean-up pass, set up a daily review of the "Needs fixing" tab as part of your routine to catch new products with issues as they come in.