Feed Title Generator: Free Google Merchant Center Title Prompt + Feed Template

What you get

Two files: the feed template as a Google Sheets copy, and the prompt as a Google Doc copy.

1. The feed template: a 10-column product sheet. The same input format we use when rewriting client feeds. You fill it from a normal Merchant Center or Shopify export, one row per product. Copy the Google Sheet

2. The prompt: the strategist system that writes the titles. A long form prompt that makes the model derive the primary keyword per SKU, pick one of 7 title templates, front load the first 70 characters, and return strict TSV. It is forbidden from inventing specs, adding promo words, or using anything not confirmed by your row or product page. Copy the Google Doc

This resource pairs with our Google Shopping title optimization page: that page teaches the rules and templates one title at a time, this one runs them across a whole batch. Both sit in our worksheet templates library.

Here is what the feed template carries and why each column exists:

Why default feed titles cost you money

Most feeds run on whatever title the platform exported: the internal product name, sometimes with a color code, rarely with the search terms buyers type. Three problems follow.

First, truncation. A product title can be up to 150 characters, but shoppers usually see only about the first 70 in ad units. If the size, pack count, or model number sits at the end, it is invisible exactly where the click decision happens.

Second, matching. Shopping and Performance Max match your products to queries largely from your feed data. A title that says Ambassador Bed competes for almost nothing. A title that names the product type, material, and size in market language competes for every query those words appear in.

Third, variant blur. When five color variants share one title, they collide with each other in matching and look identical in results.

This is rarely a small cleanup. On ThatBlanket.co, a US personalized photo blanket brand where we rebuilt the Google Ads growth system, one product export showed 1,358 products all marked Not eligible in that report, most with no campaign advertising them. That did not automatically mean Merchant Center was broken, but it showed how far feed and product structure can drift while campaigns keep spending. Feed segmentation and eligibility fixes came before any scaling decision. Titles are the same category of work: unglamorous feed hygiene that decides what your budget is even allowed to match.

What the prompt enforces that a plain AI ask does not

Anyone can paste products into a chatbot and ask for better titles. The output usually fails Merchant Center review or quietly invents specs. The prompt exists to prevent both. It enforces:

• A no invention rule. Only facts from your row, or from the product page if you allow lookup. Unconfirmed specs are omitted and flagged, never guessed.
• Google policy limits. 1 to 150 characters, no promotional text, no price, no ALL CAPS outside model codes, manufacturer brand only, clean pipe separators.
• First 70 priority. Product type, the meaning changing modifier, and the primary variant or quantity must land in the first 70 characters, then confirmed specs extend the title toward 115 to 150 characters in a fixed order.
• One template per row. Seven templates cover brand led, product led, spec led, compatibility led, pack led, ingredient led, and style led products. The model must pick exactly one and report which.
• Variant separation. Rows that are variants of one product must differ by their variant value or get flagged as a duplicate risk.
• Market awareness. UK rows get UK spelling and size naming, US rows get US conventions, and units are never converted.
• Auditable output. Strict TSV, one row out per row in, with a risk flags column (thin specs, promo terms found, title versus page mismatch, low confidence keywords) so you know exactly which rows need a human look before upload.

Every batch still gets a human pass before it touches your feed. The flags column is the review list.

How to run it

1. Step 1

   Copy both files. Save the Google Sheet template and the Google Doc prompt to your own Drive using the links above.

2. Step 2

   Fill the sheet from your feed. Export products from Merchant Center or your store platform and map them into the 10 columns, one row per product. Paste the landing page URL in Link if you want the model to verify specs.

3. Step 3

   Set the run settings. At the top of the prompt, set your market, whether web lookup is allowed, brand demand, and compliance mode. Defaults are already sensible.

4. Step 4

   Run a batch of 25 to 30 rows. Paste the prompt into a capable AI model, attach or paste your rows, and let it return the TSV. Paste the block into your sheet and confirm the row count matches.

5. Step 5

   Review flags, then upload. Check every flagged row against its landing page, then apply the approved titles through a supplemental feed or your platform's feed fields.

One warning from client use: the output is only as good as the sheet. Thin rows produce short titles, and the prompt will tell you so with a short_title flag naming the missing specs. That flag list is your data collection to-do, not a defect.

Where this fits at real spend

If Shopping is a few hundred dollars a month for you, run the free files yourself. Nothing here is held back.

At $10,000 or more per month, titles stop being copywriting and become budget routing. The queries your titles can match are the queries your money can enter. That is why title work sits inside the feed stage of every Google Shopping engagement we run, next to feed segmentation, product exclusions, and eligibility fixes, and inside the account structure work on our broader Google Ads management side.

When this is not the fit: catalogs of many thousands of SKUs with daily price and stock churn belong in a rules based feed tool, with AI batches reserved for the revenue driving subset. And if tracking or landing pages are broken, fix those first. Better titles pointed at a page that cannot convert just measure the problem faster.

Done for you for $99

If you would rather not build the sheet and run the batches, we do it for $99 for a batch of up to 100 SKUs.

You send your feed export and store URL through the form. We map it into the template, run the full system with web lookup on, review every flag against your landing pages, and send back the finished sheet: final title, two alternates, template used, character counts, and the flag notes for anything your product data should add. You upload it, or we hand it to whoever manages your feed.

It is the same title workflow we run inside client accounts spending $10,000+/month in the US and £5,000+/month in the UK on Google Ads. You get the deliverable without the retainer.