The fix queue: what to fix first
The ranked list of problems under your health score. How it groups issues, why errors come first, what the product counts mean, and how to clear them.
What the fix queue is
The fix queue is the ranked list of problems shown under your health score in the Feed health panel, labelled "Fix these first." Instead of dumping every failing product on you, it groups the problems by what is wrong and orders them so the highest-impact fix is at the top.
Each row is one type of problem: a field plus a severity. For example "Missing required field title" is one row, no matter how many products are missing a title. The row tells you the plain-language reason, how many products it affects, and which channels it affects.
How the queue is ranked
The queue sorts by two rules, in order. First, errors come before warnings, because errors get products rejected and warnings do not. Second, within the same severity, the issue affecting the most products comes first.
So the top of the queue is always the error hitting the largest number of products, and the bottom is the warning hitting the fewest. Working top to bottom recovers the most products for the least effort.
The panel shows the top 8 rows. Clearing the ones at the top usually lifts the health score the most, since they cover the most products.
- Errors are always ranked above warnings.
- Within a severity, higher affected-product count ranks first.
- Each row shows the count and the channels it appears on.
- The panel displays the top 8 issues.
How counts and channels work across feeds
The fix queue is built from the full validation breakdown of every channel feed, not a sample. If the same problem (say, a missing image) shows up on both your Google and Meta feeds, it is merged into one row and the affected channels are listed together.
The product count on a row is the total across the channels where that problem appears. That is why a single missing field can show a count larger than your product total: it is being counted once per channel feed it breaks.
Because the queue reads the stored breakdown from the latest build of each feed, it stays accurate without you re-running anything: it updates whenever a feed rebuilds.
Working the queue
Most fixes happen upstream of the export, in your field mappings or transform rules, because that is where a product's values come from. A "Missing required field" row almost always means the channel column is not mapped to a source column, or the source column is empty.
After you change a mapping or rule, rebuild the feed (or wait for the next scheduled refresh) so the queue and score recompute against the new output.
- 1Start at the top row (the biggest error) and read the plain-language reason and the field it names.
- 2Open your field mappings or transform rules and find that field.
- 3Map it to the right source column, or add a rule that fills it (for example, set identifier_exists to no for handmade items with no GTIN).
- 4Rebuild the feed, or let the next scheduled run rebuild it.
- 5Re-open the Feed health panel and confirm the row's count dropped and the score rose.
- 6Move to the next row.
AI suggestions
Next to "Fix these first" there is an AI suggestions button. On instances where it is enabled, it reads your current issues and proposes concrete fixes in plain language. If the button reports that AI suggestions are not enabled on this instance, the feature is simply turned off there; the rest of the fix queue still works exactly the same.