Product feed optimization for a Shopify brand, with AI

Product feed optimization means improving the product data a store sends to Google Merchant Center and other channels, so products are approved, match more searches and show the right details. For a Shopify brand, the feed usually starts as a straight copy of the store catalog through Shopify's Google and YouTube app. Optimizing it means rewriting titles for how people search, filling in missing attributes such as color, size and GTIN, fixing disapprovals and keeping it all in sync as products change. AI now does much of the rewriting. This answer sets out which fields matter, what the tools cost at published prices and when an automated AI workflow beats a feed tool. It is from Brand Media Works, an AI consulting and marketing data engineering firm for ecommerce brands.

What are the key fields to optimize in a Merchant Center feed?

Google requires an ID, title, description, link, image link, availability and price for every product (read 1 November 2026). Five fields decide most of a feed's performance:

1. Title. Up to 150 characters. Lead with what the shopper searches for: brand, product type and the attribute that matters most, such as size, material or color.
2. Description. Up to 5,000 characters. Plain facts about the product, not marketing copy.
3. Product category and type. Google accepts one product category and uses only the first product type value, up to 750 characters, for bidding and reporting in Shopping campaigns.
4. GTIN, MPN and brand. Google strongly recommends a GTIN where one exists, requires an MPN when there is no GTIN, and requires a brand for all new products except movies, books and music.
5. Images. Google's new minimum is 500 x 500 pixels, with enforcement beginning 31 January 2027.

Apparel has extra rules. Color, gender and age group are required for all apparel products targeted to the US, the UK, Brazil, France, Germany and Japan.

What does feed optimization cost, and when is an automated AI workflow worth building?

There are three routes, at published prices:

• Shopify's Google and YouTube app. Free to install, built by Google, and syncs products, deals and shipping to Merchant Center. It is rated 4.5 from 5,584 reviews, though 12% of ratings are one star (read 1 November 2026). It copies the catalog as it is, so the feed is only as good as the store data.
• A feed tool. DataFeedWatch lists $64 a month for 1,000 SKUs and 3 feeds, $84 for 5,000 SKUs and 10 feeds, and $239 for 30,000 SKUs. GoDataFeed recommends its Plus plan at $99 a month for 1,000 SKUs, adds $5 per extra 1,000 SKUs and $29 a month per extra ad channel, and includes AI titles and descriptions (read 1 November 2026).
• An automated AI workflow. A workflow reads the catalog, asks a model to rewrite titles and fill attributes under the brand's rules, checks the output and writes it back. Model fees are small: Amazon Bedrock lists Gemma 4 26B at $0.13 per million input tokens and $0.40 per million output tokens (read 15 October 2026), so 10,000 products at 1,000 input and 300 output tokens each costs about $2.50 per full pass (our arithmetic). n8n's hosted Starter plan, which can run the workflow, costs $24 a month (read 26 October 2026).

A feed tool is the better choice when a brand sends one catalog to many channels and needs standard rules. A custom workflow pays off when titles and attributes need the brand's own rules and an approval step, when attributes have to come from supplier files or a product information system, or when the catalog changes daily. AI workflow automation describes how Brand Media Works builds feed workflows, and Gemini API pricing in 2026 prices a 10,000-product catalog run on Google's models.

What is a product feed?

A product feed is a structured file, or an automatic sync, that lists every product a store sells with its attributes: ID, title, description, price, availability, images, brand, identifiers and category. Google Merchant Center uses it for Shopping ads and free listings, and other channels such as Meta, TikTok and Microsoft use their own versions. When the feed is wrong, products are disapproved or shown for the wrong searches, even if the store pages are correct.

How does AI improve a product feed?

• Rewriting titles in a consistent pattern, such as brand, product type, key attribute and size, within the 150-character limit.
• Filling missing attributes such as color, material, gender and age group from descriptions, images or supplier data.
• Classifying products into Google's product categories and the brand's own product types.
• Checking before sending: catching titles over the limit, missing identifiers or images below the minimum before Merchant Center disapproves them.

The check matters most. A model will occasionally invent an attribute, so a good workflow compares its output with the source data and sends uncertain changes to a person. Google's own guidance is to provide a GTIN only when it is certain to be correct.

Book a discovery call

Frequently asked questions

Do I need a feed tool if I use Shopify's Google and YouTube app?

Not always. The app syncs the catalog to Merchant Center for free. A feed tool or an AI workflow adds value when titles need rewriting, attributes are missing or the brand sends products to several channels.

How long can a Google Shopping title be?

Up to 150 characters, under Google's product data specification. Put the words shoppers search for first.

What happens on 31 January 2027?

Google begins enforcing its new product image requirements, including a minimum of 500 x 500 pixels. Brands with older or small images should check their catalog before then.