ECOMMERCE

Shopping feed title optimization

Use when product titles in your shopping feed are truncating or missing the terms buyers search.

feed-title-optimisation.md

Download .md

Optimize shopping feed titles for {{RETAILER_OR_CHANNEL}}.

Products with current titles and attributes:
{{PRODUCT_TITLES_AND_ATTRIBUTES}}

Attribute order preferred for this category: {{PREFERRED_ATTRIBUTE_ORDER}}
Search terms these products should match: {{TARGET_TERMS}}
Character limit: {{CHARACTER_LIMIT}}

Output a table: SKU | Current title | Proposed title | Characters | What changed and why | Risk.

Rules:
- Proposed titles must fit the character limit. Count characters and show the count.
- Front load the attributes most likely to match a query. The brand goes first only if the category is brand led.
- Use only attributes present in the input. Do not add a size, color, material or model number you were not given.
- No promotional text, no all caps words, no repeated words, no punctuation used as decoration.
- Mark Risk as high where the change could reduce match on an existing converting term.
- After the table, list any attribute you needed and did not have, per SKU.

Fill in before running

Replace each placeholder with your own detail. The more specific you are, the less the model invents.

{{RETAILER_OR_CHANNEL}}
{{PRODUCT_TITLES_AND_ATTRIBUTES}}
{{PREFERRED_ATTRIBUTE_ORDER}}
{{TARGET_TERMS}}
{{CHARACTER_LIMIT}}

Getting a better result

Do not batch change high spend SKUs. Run them as a separate group and watch impressions for a fortnight.
The risk column matters more than the proposed title - it tells you what not to ship.
Feed the actual search terms report rather than a keyword list. Titles should match real queries.

Questions about this prompt

When should I rewrite titles rather than change bids?

When titles are truncating in the channel or missing the words buyers type. Bidding is the usual first move and it is more expensive, and it does nothing when the title never matches the query in the first place. Do this before touching bids, and keep the scope to titles.

What do I need before rewriting the titles?

Current titles with their attributes, the attribute order that suits the category, the channel character limit, and target terms taken from the actual search terms report rather than a keyword list. A keyword list gives you titles matched to what you wish people searched for, which is not the same thing.

What does the table give me?

A table of proposed titles with character counts, what changed and why, and a risk rating, then the attributes you were missing per SKU. The risk column is worth more than the titles themselves, because it marks the changes that could lose match on a term that is currently converting.

How should I roll the new titles out?

Pushing the whole table live in one upload. Split high spend SKUs into their own group and watch impressions for a fortnight, since a title change resets what the channel has learned to match you against. And never let a size, colour or model number into a title that was not in the input.