Product Feed Quality Index 2026: We Audited 12,375 Products Across 251 Stores
We measured 12,375 products across 251 real Merchant Center accounts. Most feeds fail the same way: no identifiers, oversized titles, empty attributes.
- 12,000+PMax campaigns audited
- 200+Live ecom clients
- €200M+Tracked sales
Everyone says "optimize your feed." Almost nobody shows you what feeds actually look like. So we measured ours. All of them.
This page is the first public edition of our Product Feed Quality Index: a recurring audit of every Merchant Center account connected to our MCC. No survey, no estimates. These are field-by-field counts from real feeds, pulled on July 28, 2026.
Products sampled
12,375
Stores measured
251
No GTIN or MPN
81.3%
Titles past 80 chars
68%
The Identifier Hole Is Bigger Than Anyone Admits
An identifier is how Google knows what your product IS. With a GTIN or an MPN, Google matches your offer against its own product catalog and puts you in every auction that product belongs in. Without one, Google reads your title and description and guesses.
Guessing costs you twice. Your product enters fewer auctions, and the auctions it does enter are matched off weaker signals. This is one of the first things we fix on a new account, and it is the least glamorous work in all of Google Ads. Nobody posts a screenshot of themselves filling in MPN columns. It just quietly prints.
If your products have real barcodes, put them in. If they do not, the brand plus MPN route keeps you eligible. What you should never do is invent GTINs to fill the column: Google's Shopping ads policies treat inaccurate product data as a violation, and it is a fast path to a Merchant Center suspension.
How Long Should a Google Shopping Product Title Be?
Google's product data specification allows 150 characters in the title, and apps love to use the room: stuffing color, size, marketing phrases, and the store name into one long string. Not one sampled title broke the 150 cap, because Merchant Center rejects those outright. But the display cut sits far lower, at roughly 70 characters depending on placement.
The other direction is nearly empty: only 4.2% of titles are under 50 characters. Short titles are rare because apps auto-build them from product names plus variants. The real problem is order, not length. If the words that sell the product sit at position 90, they may as well not exist.
Front-load the first 70 characters: product type, brand, the attribute buyers search for, the color. Everything after that is for the algorithm, not the shopper. The full title playbook is in our feed optimization guide.
The Attributes That Separate Stores Are Almost Universally Empty
Here is the pattern that jumped out of the data. Sort the attributes by how often they are filled and you get two clean groups:
| Fill rate | Verdict | |
|---|---|---|
| Brand | 97.5% | Everyone fills it |
| Condition | 96.2% | Everyone fills it |
| Color | 84.9% | Most fill it |
| Product type | 80.8% | Most fill it |
| Size | 80.1% | Most fill it |
| 2+ extra images | 41% | Half the fleet, thin |
| Product details | 10.8% | Almost nobody |
| GTIN or MPN | 18.7% | Almost nobody |
| Material or pattern | 5% | Almost nobody |
The top group is what Shopify and feed apps fill automatically. The bottom group is what a human has to add on purpose. That is the whole story of feed quality in one table: the automatic fields are done everywhere, so they no longer separate anyone. The manual fields are empty everywhere, so filling them puts you in the top tenth of feeds by that attribute on day one.
Images tell the same story. Only 41% of products carry two or more additional images beyond the main photo. Google pulls extra images into Shopping formats and AI results, and most stores give it exactly one photo to work with.
What Should You Fix in Your Feed First?
Identifiers, then titles, then one manual attribute across the whole catalog. Three moves, in order of payoff per hour:
- Fill identifiers first. GTIN where a real barcode exists, brand plus MPN where it does not. This unlocks auctions everything else builds on.
- Rewrite titles front-first. The selling words go in the first 70 characters. Test against the display, not the character counter.
- Fill one manual attribute across the whole catalog. Material for fashion and furniture, pattern where it applies, product_detail for specs. One pass over the catalog beats a hundred bid tweaks.
We run this exact sequence on client feeds through HyperFeed, and it is step four of the free learn path. If you run a dropshipping store, the full context lives in the complete Google Ads for dropshipping guide.
Methodology
The claims above come from one snapshot, and here is exactly what it is.
- Source. The nightly Merchant Center snapshot across every account connected to the ZenoX MCC, taken July 28, 2026.
- Sample. Up to 50 products per store, from the first page of the feed. 12,375 products across 251 stores with at least one sampled product.
- Measured at arrival. Many accounts are measured at onboarding, before our optimization work starts. These numbers describe how feeds arrive, built by apps on default settings, not how they leave.
- Boundaries. "Past 80 characters" uses our snapshot's real bucket boundary at 80. Google's roughly 70-character display cut is a separate, well-known display behavior, and we state it separately on purpose.
- What we left out. Disapproval-reason data exists for only 10 accounts in this snapshot, so we do not publish it. Approval-rate totals in the source were internally inconsistent, so we do not publish those either. No by-niche breakdowns: our store labeler tags 81.9% of stores as Fashion, which makes niche cuts statistically dishonest.
- Cadence. We re-measure on a recurring basis. The numbers on this page update when the index does, with the snapshot date shown at the top.
This index sits alongside the rest of our original research, built the same way: our own portfolio data, counted rather than surveyed.
Free playbooks
Fix Your Own Feed with the Same Playbooks
The SOPs we use to close exactly the gaps this study measures. Free on Notion, no email gate.
- FeedsFeed Optimization MasteryThe deep version. Every feed attribute that moves clicks and sales, one by one.
- FeedsFeed Optimization GuideTitles, labels, and the feed work that drives most of the lift on an ecom account.
- FeedsGoogle Shopping Feed OptimizationHow to write product titles that match what people search, so your ads show more.
- FeedsCompetitive Price MonitoringPrice is a ranking signal in Shopping. How to watch rivals and stay competitive.
- ScalingGoogle Shopping Profit OptimizationsShopping accounts leak profit in a hundred small places - wrong products getting budget, bad signals. Plug the leaks.
- Merchant CenterBenefits of Switching Your CSS PartnerA CSS switch can cut your Shopping costs. What it is and why it helps.
Questions about the data? Ask in the free community and I will pull the numbers.
Frequently Asked Questions
What is the Product Feed Quality Index?
It is our recurring audit of real Merchant Center feeds. We sample up to 50 products per store across every account connected to our MCC and measure the fields that decide Shopping performance: identifiers, titles, descriptions, images, and attributes. This edition covers 12,375 products across 251 stores, measured on July 28, 2026.
Why do GTINs and MPNs matter for Google Shopping?
Identifiers tell Google exactly which product you sell. With a GTIN or MPN, Google can match your offer to its product catalog, compare it against other sellers, and show it in more places. Without one, Google has to guess from your title and description, and your products enter fewer auctions. Most stores skip both fields entirely.
How long should a Google Shopping product title be?
Merchant Center accepts up to 150 characters, but most Shopping placements cut the display at around 70. Put the words that sell the product in the first 70 characters: product type, brand, key attribute, color. In our data, 68% of titles run past 80 characters, which means the tail of most titles is invisible in the ad.
Are these numbers from stores you optimized?
No, mostly the opposite. Many accounts in the sample are measured at onboarding, before our feed work starts. That is the point: this is what feeds look like when they arrive, built by apps and themes on default settings. The gap between these numbers and a worked feed is the opportunity.
Can I see the methodology?
Yes, it is spelled out at the bottom of this page. Snapshot date, sample size per store, how each metric is computed, and the stats we chose NOT to publish because the underlying data was too thin to stand behind.
