Product Feed Optimization · focused work
Product feed title optimization
The title in a product feed is the strongest matching signal a shopping system has and the first thing a shopper reads in a listing. It is not the page title and it is usually not the name printed on the product page. It is a compressed description assembled from the words buyers actually type, in the order they type them.
This is the right work if
- Feed titles are copied from product page names written for the brand rather than for search
- Listings appear for the wrong size, colour or model because the title never says which
- Clothing titles omit who the item is for, its colour or its size while competing listings state all three
- Search term reports show queries the products would match if the title mentioned the attribute
- Titles were rewritten once and nobody compared performance before and after
What it is
What title work actually produces
The deliverable is a title pattern for each product category, plus the mechanism that generates titles from real product attributes and keeps them current. Different categories need different structures. Clothing needs brand, who it is for, the garment, colour, size and often material. Electronics needs brand, model number and the specification people search by. Consumables need brand, product and the quantity or volume. Parts need brand, part number and the fitment that makes the listing useful. A single pattern applied across a mixed catalog produces good titles for one category and poor ones for the rest.
Order matters as much as content, because the title is truncated in most placements and only the opening words are read. The vocabulary comes from evidence rather than instinct: the account's own search term reports, the store's internal search, and the phrasing customers use in reviews and support. Policy sets the boundary, since promotional wording, block capitals and marketing claims are not permitted, so a feed title is descriptive rather than persuasive. Implementation goes through feed rules or a supplemental source so the store's own product names stay readable for shoppers, and every change is measured against a group of products left unchanged.
How it is done
The work, in order
What changes
- Titles use the words shoppers type, in the order they type them
- Listings match the correct size, colour and model instead of a near neighbour
- Product names on the site stay readable while the feed carries the search version
- Title changes are judged against an unchanged control group
Take the vocabulary from real queries
Start with the account's own search terms and the store's internal search, then supplement with research. Those two sources show the phrasings that already lead to orders for this catalog, which is more reliable than volume estimates for terms the store has never served.
Design one pattern per category
Write the attribute order for each product category as a template naming the fields it draws on, and note which are mandatory. Reviewing these templates together makes the inconsistencies obvious before any of them reach a live feed.
Put the identifying words first
Lead with what a shopper would type to find this exact item, because the visible portion of a title is short. Descriptive extras belong at the end where truncation costs nothing, and the store name rarely earns its place at the front.
Generate from attributes, not from text
Build the title from structured product fields so it stays correct when a product changes. Where an attribute is missing, fill it in the source data rather than typing a one-off title, since the manual version will not survive the next catalog update.
Stay inside the policy and the limit
No promotional phrases, no block capitals for emphasis, no claims about being cheapest or best. Keep within the character limit and treat the limit as generous space for description rather than as a target to fill with repeated words.
Apply through rules and measure
Implement in feed rules or a supplemental source so the store's data stays clean, then hold a comparable group of products back unchanged and compare impressions, clicks and conversions between the two over a full sales cycle.
Platform notes
Shopify
Feed apps typically build a variant title by appending option values to the product title, which produces workable clothing titles and poor ones for categories where the option is not what a shopper searches by, so the mapping usually needs overriding.
WooCommerce
Variation names are derived from attribute terms, so the quality of the title depends on how attributes were set up in the first place, and tidying the attribute taxonomy often improves more than editing the feed does.
Does the brand always belong at the start of the title?
It belongs first when people search by brand, which is true for recognised names and for parts where the brand determines fit. For unbranded or own-label goods, the product type is what shoppers type, so leading with a brand nobody searches for spends the most valuable words in the title.
How much of the title is actually seen?
Considerably less than the limit allows, and the visible amount differs by placement and device. Treat the opening words as the title a person reads and the remainder as detail that helps matching, and never rely on something at the end being noticed.
Can one set of titles serve every channel?
Usually as a starting point, since most shopping channels reward the same descriptive structure. Where a channel has different length limits or different policy on capitalisation and abbreviations, a per-channel supplemental source is a cleaner answer than compromising the main title.
Should the product page name change to match?
Not necessarily. The page has context the feed does not, so a shorter, more human name often reads better on site while the feed carries the fully qualified version. Where the page name is genuinely unclear to a first-time visitor, that is worth fixing on its own merits.
Related work and answers
Find the leak.
A free Growth Analysis ranks what your store should fix first, by revenue at stake.