Title Hygiene

What you get

• Three tiles: Offers, Duplicate title groups and Offers repeating a word.
• Title length: characters per title as a histogram, with the average and median, and a note when titles contain CJK text.
• Description length: characters per description over the offers that have one, and how many carry no description.
• Top words: the most frequent significant words, by the number of titles containing them. Stopwords, bare numbers and one-character tokens are skipped, and a word that appears in the offer’s brand on at least half of its titles is marked as a brand token.
• Duplicate titles: offers sharing one title after trimming, collapsing whitespace, Unicode normalisation and lowercasing, largest groups first, with up to five example offer ids each.
• Repeated words: offers that repeat a significant word inside one title, with examples.
• Feed versus page title: offers whose crawled product page carries a different title than the feed, from the latest feed audit run.

How to use it

1. Open the property, expand E-COMMERCE in the left sidebar, and click Title Hygiene.
2. Start with Duplicate titles and Repeated words: both are direct fixes in the feed.
3. Use Top words to see which terms your titles lean on, and which are only brand names.
4. Run a Feed Audit to fill Feed versus page title.

Controls

• Refresh reloads both parts of the report.
• Copy buttons on the tables, and CSV and HTML export on Top words.

Data source and freshness

• The summary reads the catalog that the Merchant Center link syncs. Until the warehouse holds the catalog, it comes from a live Merchant API read, and the page says so.
• The feed-versus-page diff is computed by the feed audit; this report only reads it, so it is as fresh as the latest completed run.

Limits

• A live Merchant API read is cut at 5,000 products, and the page says when that happened.
• When there are more duplicate groups than fit, the card says more exist than are shown.