What Is a UCP Score? Read Your Feed Health at a Glance

What a UCP Score Actually Measures

A UCP Score is a single number, 0 to 100, that tells you how ready one product in your feed is to be found and recommended by AI. One glance and you know whether that item is pulling its weight or quietly getting skipped.

It isn’t a Google metric. It’s UCP Radar’s read on how structured and AI-readable a product’s data is, and whether an assistant like ChatGPT or Gemini has enough to confidently put your item in front of a shopper who’s about to spend money. Above 80, a product is well-positioned for AI recommendation. Below 50, it’s probably being passed over for a competitor whose data is complete.

Here’s the gap the score exists to expose. Most product feeds land between 25 and 40 out of 100 on AI-readiness. When colors, sizes, materials, and real descriptions are missing, AI agents can’t match those products to a shopper’s question, so they recommend someone else. The UCP Score puts a number on that gap, product by product, so you can see it before it costs you the sale.

Think of it as a credit score for your catalog. You don’t need to read the underlying report to know where you stand. The number tells you, and it tells you per product instead of as a vague feed-wide average.

How to Read the Score: the 0–100 Bands

The score is easiest to use when you read it in bands rather than chasing one perfect number. Three ranges cover almost every product you’ll look at.

A product above 80 is in good shape. It has the description depth, the filled attributes, and the accurate availability an AI assistant needs to recommend it without guessing. These are the items already working for you on Google Shopping and, increasingly, on AI surfaces. Leave them alone and keep an eye on them.

A product between 50 and 80 is the middle of the pack. It’s eligible but not compelling. Usually a few attributes are missing, or the description is too thin to answer a real question. These are your fastest wins, because small fixes push them into recommendable territory.

A product below 50 is the one to worry about. At that level the data is too sparse for an assistant to trust, so the product is most likely being skipped. If a below-50 product is one of your revenue drivers, that’s money leaking out of the catalog every day it sits there.

The bands matter more than the exact figure. A product that climbs from 41 to 58 has crossed a real line. A product that drifts from 84 to 82 has not. Read the movement, not the decimals.

UCP Score vs. GMC Score: Two Numbers, Two Jobs

The UCP Score has a sibling, and people mix the two up. UCP Radar calculates a GMC Score for every product as well, and they answer different questions.

You need both, because they fail in different directions. A product can be perfectly compliant with Google’s rules, with a clean GMC Score, and still be invisible to AI because its description is two sentences and half its attributes are blank. Compliance gets you into the auction. Depth gets you recommended. The GMC Score keeps your Shopping campaigns running; the UCP Score decides whether AI assistants put you on the shortlist.

Reading them together is the point. A high GMC Score paired with a low UCP Score is the most common pattern we see, and it’s also the most expensive, because the merchant assumes everything is fine while AI quietly routes shoppers elsewhere.

What Moves Your UCP Score

You don’t need the exact recipe to cook with it. A handful of data-quality factors do most of the work, and they’re the same feed-quality basics that have always governed feed performance.

Description depth. A two-line marketing blurb doesn’t carry enough signal to answer a shopping question. Descriptions that say what the item is, who it’s for, and why it’s different move the score the most. This is the single biggest lever, and the one most catalogs ignore.

Attribute completeness. Color, size, material, GTIN, compatibility. Every filled attribute is another way for an AI to match your product to a specific query. A blank attribute is a match you can’t win.

Freshness and accuracy. A feed that says “in stock” when it isn’t gets a product deprioritized fast. Current, accurate data holds the score up. Stale data pulls it down.

Trust markers. Real GTINs, clean images, and reviews give an assistant the confidence to put your product ahead of an unknown one.

A high GMC Score means Google will run your ads. A high UCP Score means AI will recommend your products. Most merchants only track the first and wonder where the second went.

One reassurance worth stating plainly: raising the score doesn’t mean an AI will rewrite your brand into something you don’t recognize. Brand Protector keeps brand names, model numbers, and product identifiers locked during optimization, so “Sony WH-1000XM5” stays exactly that. It approves 98.2% of changes cleanly and flags only the 1.8% that genuinely need a human to look.

Reading Your Feed Health at a Glance

A single product’s score is useful. The real value shows up when you look at the whole catalog at once.

Sort your products by UCP Score, lowest first, and overlay revenue. Now the priority order writes itself. The low-scoring, high-revenue products at the top of that list are where a fix returns the most. You’re not trying to drag every SKU to 92. You’re trying to lift the few products that move the most money out of the danger zone first.

This is also how you catch drift. Feeds decay. Prices change, items sell out, and new products arrive barebones from your store platform. Watching the score over time tells you when feed health is slipping before it shows up as lost ROAS. A feed that averaged 78 last month and 64 this month is telling you something happened, usually a wave of new products that landed with thin data.

The fixes themselves are rarely the hard part. Knowing which 200 products out of 4,000 to fix first is the hard part, and a score per product makes that decision in seconds instead of a week of spreadsheet work.

Conclusion

A UCP Score turns a vague worry, “is my feed good enough for AI?”, into a number you can act on. Read it in bands, pair it with your GMC Score, and sort your catalog by it to see exactly which products are recommendable and which are getting skipped. Treat the score as a readout rather than a target. It shows you where the money is leaking. UCP Radar scans your existing Merchant Center feed against more than 50 validation rules, scores every product on both axes, and shows you the before-and-after, so you can connect your feed, see your scores for free, and fix the products that matter before a competitor’s data takes the recommendation that should have been yours.