Google Shopping Feed Optimisation: Your Product Titles Decide What You Show For

If your Shopping campaigns are underperforming, the bidding is rarely the problem. Google Shopping feed optimisation is where the money sits, because your product feed, not a keyword list, decides which searches you appear for. Shopping campaigns have no keywords. They have a spreadsheet of product data, and Google matches that data to queries. Get the data wrong and no amount of bid tinkering will rescue the campaign.

Most feeds we inherit have never been touched since the platform generated them. Shopify or WooCommerce exported whatever sat in the product name field, and nobody looked again. That is one of the cheapest fixes available in Google Ads, and it usually takes an afternoon rather than a retainer.

Your product title is the closest thing to keyword targeting you get

Google's documentation is direct about this. In its guidance on the title attribute, it states that your title "will be used to match your product to a customer's search" and tells you to include the important details that define the product. That one line should change how you write every title in the feed.

"The Weekender" tells Google nothing. "Waterproof Canvas Weekend Bag, 40L, Olive Green" tells it the product type, the material, the defining feature, the capacity and the colour. The second title can match dozens of realistic queries. The first can match almost none, because nobody searches for your internal product names.

Google allows 1 to 150 characters for the title and explicitly recommends using all of them. Most of the feeds we audit use a fraction of that. Every character you leave empty is matching surface you have given away for nothing.

Front-load the first 70 characters

Length is not the same as ordering. Google's own best practice notes that users will usually notice only the first 70 or fewer characters of your title, depending on screen size. So the structure matters as much as the content.

A workable order: brand, then product type, then the key attribute a buyer would search on, then variant details. What you want to avoid is brand, collection name, an adjective your copywriter liked, and then finally the product type at character 90 where nobody reads it.

For a speciality coffee retailer, that means roast level, grind type and bag weight go near the front, because those are the words people actually type. We built the storefront for Manchester coffee brand no worries, bru around branding the founders had developed themselves, and the same principle applied there as anywhere: the attributes a buyer uses to choose belong in the product data, not buried in the description.

The rules that will get you disapproved

Google's title specification sets out minimum requirements, and they are not suggestions. No promotional text in the title, which includes price, sale price, sale dates, shipping, delivery timings and your company name. No capital letters for emphasis. No gimmicky symbols or HTML tags. Use professional, grammatically correct language, and distinguish variants properly by putting the differentiating detail in the title as well as in the relevant attribute.

Two details worth knowing. If your title exceeds the character limit, Google truncates it and warns you. And if you submit both title and structured_title, Google uses title only, which catches out a lot of merchants who have started generating titles with AI and assumed the structured version takes priority.

Images disapprove products faster than anything else

Titles decide what you match. Images decide whether the product serves at all. Google requires an unobstructed view of the product, and its guidance on text on image is unambiguous: products that do not meet that requirement are disapproved and stay disapproved until the images are updated.

The trap for smaller retailers is reusing social assets in the feed. A "free UK delivery" flash or a discount roundel is fine on Instagram and fatal in Merchant Center. So are watermarks, including your own logo. If you have a large catalogue with this problem, Merchant Center's automatic image improvements can attempt fixes for you, though changes take 24 to 72 hours to show on the Needs attention page, so do not expect same-day recovery before a sale.

Write a formula, not four hundred titles

Nobody rewrites a 400-SKU catalogue by hand, and you should not try. Build a title rule per product type in your feed management tool or your platform's feed settings, along the lines of {brand} {product type} {key attribute} - {variant} - {size}. Then apply it, and hand-check the twenty products that generate most of your revenue.

That eighty-twenty split is the honest version of feed work. The tail benefits from the rule. The head benefits from someone actually reading the title and asking whether a stranger would recognise the product from it.

Feed work has a ceiling, and you should know where it is

Once your titles are specific, your images are clean, your product types are mapped and your identifiers are correct, feed optimisation stops being the constraint. At that point you are paying for qualified clicks and losing them somewhere else: a slow product page, a variant selector that hides sizes, a checkout that asks for an account before it asks for a card.

That is a different job. It belongs in conversion rate optimisation and, often enough, in web development. The mistake is spending another month on feed rules when the feed is already fine and the site is the problem. Diagnose in that order, because the feed is cheaper to fix and the site is more expensive to ignore.

If your Shopping campaigns are spending without much to show for it, we will look at the feed before we touch the bids. Get in touch and tell us what you are selling.