Google Shopping Feed Attributes: Required, Optional, and What Changed in 2026

Google's product data specification runs to dozens of attributes, and the documentation gives every one of them the same weight. That is useful as a reference and unhelpful as a plan. Somebody still has to decide what to populate first, and the spec does not rank anything for you.

So here is the ranking. Seven attributes are required for every product you submit. A second group is required only when it applies to your product, and that group causes most of the disapprovals we see. Everything else is optional, which is the most misread word in the whole specification. Below is what sits in each tier, the format rules that trip people up, and the three things Google changed in 2026.

The seven attributes every product must have

Miss one of these and the product does not serve. There is no partial credit.

• id: your unique identifier for the product, up to 50 characters. Keep it stable, because Google keys everything else to it.
• title: the product name, up to 150 characters.
• description: up to 5,000 characters.
• link: the product's landing page on your site. It has to start with http or https.
• image_link: the main product image. The resolution rules changed this year, see below.
• price: in ISO 4217 currency format.
• availability: one of in_stock, out_of_stock, preorder, or backorder.

Look at what is absent from that list. No brand, no gtin, no google_product_category, no product_type. Those live in the next two tiers, and that is the whole reason a feed can pass every required check and still perform badly.

Required if applicable, the tier that causes disapprovals

These attributes are not required for every product, so they are easy to skip. Then one slice of the catalog gets disapproved for a rule that only ever applied to that slice, and the report lands on somebody who did not build the feed.

• brand: required for all new products, except movies, books, and musical recordings.
• gtin: strongly recommended whenever the manufacturer assigned one. Google's guidance on accuracy is blunt: "Only provide a GTIN if you are sure it is correct. When in doubt do not provide a GTIN." A wrong GTIN does more damage than a missing one.
• mpn: needed when the product has no manufacturer-assigned GTIN.
• identifier_exists: set this to no only when you are certain the product has no GTIN, MPN, or brand. Custom apparel, handmade goods, vintage pieces. Google's own instruction is to use it "for products that don't have a GTIN, MPN, or brand," and it will warn you if it finds evidence that identifiers exist after all.
• condition: required when the product is used or refurbished.
• availability_date: required when availability is set to preorder.
• age_group, gender, color, size: required for apparel in select countries. For free listings, age_group, gender, and color are required across all Apparel & Accessories.
• item_group_id: required for variants in select countries, up to 50 characters. It is the key that ties a variant group together.
• is_bundle and multipack: required for bundles and multipacks in select countries.

The country qualifier is the trap. "Required in select countries" reads as optional right up until you learn you are selling in one of them.

The optional attributes, where the competition actually happens

Optional means Google will not disapprove your product for leaving the field blank. It does not mean the field is decorative.

Across the catalogs we work on, the required block is almost never the problem. Feeds clear every required check and still underperform, because the optional block is empty. The required attributes decide whether a product is eligible to appear. The optional ones influence how well it competes once it is there, and they carry most of what a shopper, or a machine choosing between three near-identical products, needs to tell them apart.

The clearest evidence is Google's own newest fields. Every one of the six conversational attributes is optional: question_and_answer, document_link, related_product, item_group_title, variant_option, and popularity_rank. Those are the fields AI shopping assistants lean on hardest, and not one of them is mandatory. We break down the formatting for each in our field-by-field setup guide.

The optional fields worth your attention first:

• google_product_category and product_type: the first is Google's taxonomy, the second is yours. product_type accepts up to 750 characters and is the one you control.
• additional_image_link: up to 10 more images per product.
• sale_price and sale_price_effective_date: the pair that makes a promotional price render correctly instead of just being a lower price.
• product_highlight: short benefit statements, up to 150 characters each, between 2 and 100 of them.
• product_detail: structured specifications as name and value pairs, up to 1,000 characters per value.
• material and pattern: 200 and 100 characters respectively. Cheap to populate, and they answer real shopper questions.
• custom_label_0 through custom_label_4: 100 characters each, invisible to shoppers, and the backbone of any serious campaign segmentation.

What changed in the 2026 spec update

Google published a 2026 update to the product data specification. Three changes are worth acting on.

Images move to a 500 by 500 pixel minimum. The minimum resolution for image_link and additional_image_link rises to 500 by 500 pixels across all product categories. Google started issuing warnings on April 14, 2026, and enforcement begins January 31, 2027. Google has said it will automatically optimize some undersized images, though that is not a plan you want to depend on. If your catalog still carries legacy thumbnails, you have until the end of January 2027, and reshooting or re-rendering images at scale always takes longer than anyone budgets for.

video_link is live. A new optional attribute for a product video. Google began reporting technical validation errors on April 14, 2026, and stated that "serving as well as policy and quality validation will not begin until June 30, 2026." That date has passed, so a video you submit today is subject to policy and quality validation and is eligible to serve. A policy or quality error on the video stops the video from serving; it does not take down the product offer.

New shipping sub-attributes. handling_cutoff_time sets the daily deadline for processing orders, and minimum_order_value specifies the minimum spend required for purchase and shipping. Two more, loyalty_program_label and loyalty_tier_label, sit under shipping for merchants running loyalty benefits. If you compete on delivery speed, handling_cutoff_time is the one to look at first.

How to audit a feed against the spec

An audit against the entire specification produces a list nobody can action. Score against the three tiers instead, in this order:

• Required coverage, which should be exactly 100 percent. Anything less is lost inventory, and it is the only tier where the target is not a judgment call.
• Conditional rules that apply to you. Do you sell apparel, variants, used goods, preorders, bundles? Each yes turns on a rule. Segment the catalog and check each segment separately, because a catalog-wide average will hide a fully disapproved segment.
• Optional coverage per attribute, ranked by revenue. Report it as a percentage of SKUs, attribute by attribute, and start with your best sellers rather than the top of an alphabetical list.

One caution that costs people real money: coverage is not correctness. A field populated with a confidently wrong value passes every coverage report you will ever run and fails at Google. That is the argument for validating values against a real source, and for testing changes through a supplemental feed, which layers new columns on top of your primary feed without touching it and can be removed if something looks wrong.

Where to start

Get required to 100 percent. Resolve every conditional rule that touches your catalog. Then work optional coverage down from your best sellers. That order holds because the first two tiers protect revenue you already have, and the third is where new revenue comes from.

The third tier is the hard one, since filling it well is a research problem repeated once per SKU. That is what our Merchant Center Enrichment service is built for: AI agents that ground every value in a real source on your own site, validate it, and publish through a supplemental feed with a human approving the work. For the wider picture on the platform, our Google Merchant Center overview is the place to start.

Want to know where your feed really stands against the spec? Book a call and we will run a coverage read on your catalog.