By Tom Kató & László Bali · ppcout.com · 15 minute read
Most e-commerce advertisers obsess over bids, budgets, and campaign settings — and ignore the one asset that decides whether any of it works: the product feed. In Shopping and Performance Max campaigns, your feed is your targeting. Google doesn't match your products to searches based on keywords you choose; it matches them based on the titles, descriptions, and attributes you send. A weak feed means your ads show for the wrong searches, miss the right ones, and pay more for both.
The frustrating part is that feed problems are invisible in the campaign interface. Your dashboard shows impressions, clicks, and a ROAS number — it never shows you the high-intent searches you're not eligible for because a product title says "XZ-400 Blue" instead of what customers actually type. The account can look perfectly healthy while a meaningful share of your catalog is invisible, mispositioned, or funding the wrong auctions.
This guide covers how Google actually uses your feed, the fields that matter most, the mistakes that quietly cap your revenue, and how to turn the feed into a durable competitive advantage. If you'd like to know how much your current feed is costing you, a professional Google Ads audit includes a full feed review alongside your campaign structure and tracking.
In a Search campaign, you tell Google which searches you want with keywords. In Shopping and Performance Max, that control is gone — Google reads your feed and decides which searches your products are relevant for. Every word in your title, every attribute you fill in or leave blank, is a signal that expands or shrinks the set of auctions you can enter. There is no keyword tab to fix later; the feed is the targeting layer, and everything downstream inherits its quality.
This means feed optimization isn't a technical chore; it's targeting strategy. Two stores selling the identical product can see completely different results simply because one describes it the way customers search and the other describes it the way the warehouse labels it. The algorithm can't guess what you didn't tell it — and it won't warn you about the auctions you never entered.
Low impression volume, irrelevant search terms, weak ROAS on certain products — these look like campaign issues, so advertisers respond with bid changes and budget shifts. But if the feed is telling Google the wrong story, no campaign setting can compensate. Raising a bid on a poorly described product just means paying more to appear in the same wrong auctions. Before touching bids, the honest first question is always: does the feed describe these products the way real customers search for them?
A useful diagnostic habit is to work backwards from the search terms report. If your Shopping ads keep matching to queries that have nothing to do with the product, the feed is over-claiming relevance somewhere — usually a vague title or a missing product type. If high-intent queries you know exist never show up at all, the feed is under-claiming: the words customers use simply aren't in your product data.
Feed quality doesn't just decide eligibility — it feeds a loop. Better product data earns more relevant impressions; relevant impressions earn higher click-through rates; higher engagement gives the algorithm cleaner conversion data; cleaner data lets smart bidding position you more competitively for the searches that actually convert. Each turn of that loop strengthens the next. The reverse is also true: a weak feed produces noisy data, which degrades bidding, which buys worse traffic, which produces noisier data. Feed work is upstream of everything, which is why it belongs at the start of any optimization effort, not the end.
The title is the strongest relevance signal in the feed, and Google weighs the first words most heavily. Yet most stores export titles straight from their shop system: internal model numbers, brand-first ordering for products nobody searches by brand, or truncated names that omit what the product actually is.
A strong title front-loads what customers search for: brand, product type, and key attributes like gender, color, size, or material — in an order that mirrors real queries. "Nike Air Zoom Pegasus 40 Men's Running Shoes, Black, Size 44" will enter dramatically more relevant auctions than "Pegasus 40 BLK-44." Same product, same price, same landing page — completely different visibility.
There's no single correct title formula, because categories are searched differently. For apparel, attributes dominate: brand + gender + product type + color + size covers most real queries. For electronics, model numbers genuinely matter — customers search "Sony WH-1000XM5" — so brand + model + product type + key spec is the right order. For home goods and furniture, descriptive attributes like material, dimensions, and style ("oak", "160cm", "Scandinavian") carry the intent. For consumables, size and count ("500ml", "3-pack") are often the deciding words. The principle is constant even as the formula changes: put the words your customers actually type, in roughly the order they type them, as early in the title as possible.
Better titles mean more relevant impressions, which mean higher click-through rates, which feed the algorithm better data, which improves how competitively you can bid. It's the rare change that lifts volume and efficiency at the same time — without spending an additional cent. Start with your best sellers: the products already earning most of your budget are where title improvements pay back fastest, and the search terms report for those products tells you exactly which customer vocabulary is missing.
Treat title changes as tests, not decrees. Change titles on a defined product set, note the date, and watch impressions, CTR, and conversions over the following weeks against a comparable control group. Because titles affect matching, expect the search term mix to shift — that's the point — and judge the change on whether the new traffic converts better, not just on whether volume moved.
Beyond the title, a set of attributes decides how often Google trusts your product enough to show it: GTIN, brand, product type, Google product category, color, size, gender, and material. Each empty field is a missed matching opportunity — and in some cases, a limitation Google applies quietly rather than flagging as an error.
GTINs deserve special attention. They let Google match your product to its full catalog knowledge — reviews, comparison data, competitive benchmarks, and the aggregated performance history of that exact product across the web. Products with correct GTINs consistently see better performance, and missing or wrong ones can restrict your eligibility without any warning in your dashboard. Wherever a barcode exists, submit it; for genuinely custom or own-brand products without barcodes, set the identifier-exists flag correctly instead of leaving the field blank or, worse, inventing values — fabricated GTINs are a fast route to item disapprovals.
The product_type field is your own categorization and doubly valuable: it improves matching and gives you a clean way to segment campaigns later — by margin, by category, by season. Use a full hierarchical path ("Home > Furniture > Chairs > Office Chairs") rather than a single word, and keep it consistent across the catalog. Stores that fill it in thoughtfully find campaign structure much easier down the road, because the feed already contains the dimensions the account needs to be organized around.
Descriptions matter more than most advertisers assume — they're a secondary matching surface where you can naturally include the use cases, synonyms, and long-tail vocabulary that don't fit in the title. Write them for customers, front-load the important content, and avoid pasting in your homepage boilerplate. Images are a click-through lever: clean, guideline-compliant product photos on white or lifestyle backgrounds, without watermarks or promotional overlays, and with the correct variant image per color. And price accuracy is a trust signal — Google compares your feed price against your landing page and against competitors, so a feed that lags your site's price changes erodes both eligibility and click-through.
Custom labels (custom_label_0 through 4) don't affect matching at all — Google's algorithm doesn't read them for relevance. They exist purely so you can slice the catalog inside campaigns: by margin band, price bracket, performance tier, seasonality, or stock depth. That makes them the bridge between feed and strategy. A feed with thoughtful custom labels lets you give high-margin products their own budget and target, quarantine clearance items, and push seasonal winners at the right moment — none of which is possible when the whole catalog is one undifferentiated blob.
The most expensive feed problems don't produce error messages. They produce an account that looks healthy — spending its budget, showing a plausible ROAS — while a share of the catalog is invisible or mispositioned. Four failure modes account for most of the damage.
Price or availability mismatches between your feed and your site erode trust and trigger disapprovals. A feed that updates once a day while your prices change hourly means you spend part of every day advertising mispriced or out-of-stock items — paying for clicks that land on a page that breaks the promise the ad made. During sales periods, when prices move fastest and traffic is most valuable, this gap is at its widest and most expensive. Sync frequency should match your price and stock volatility, not your platform's default.
Images that violate guidelines — watermarks, promotional text overlays, borders, placeholder graphics — reduce serving quietly, and often only for the affected items, so the aggregate numbers hide it. Variant mismatches are the subtler version: a black shoe shown with the red variant's photo doesn't get disapproved, it just converts worse. Image audits belong in the feed routine, not just the disapproval queue.
Then there's the mistake that costs the most: sending your entire catalog into one campaign with one target. Your feed contains high-margin winners and low-margin traffic products, and treating them identically means the algorithm optimizes for revenue wherever it's cheapest — usually your least profitable items. The blended ROAS looks acceptable while your best products are starved and your worst ones overfunded. Feed-based segmentation with custom labels — by margin, price bracket, or performance tier — is how you take back control: the labels live in the feed, so the campaign structure can follow value instead of averaging over it.
Every catalog accumulates disapproved and "limited" items over time — policy flags, missing identifiers, mismatch warnings. Individually each one is small; collectively they can quietly remove a double-digit percentage of your catalog from auctions. Because Merchant Center buries this in a diagnostics tab most advertisers rarely open, the attrition goes unnoticed for months. A monthly pass through item-level diagnostics, prioritized by the revenue the affected products used to generate, catches the leak before it compounds. The pattern across all four failure modes is the same: the interface looks fine, and the gap between "looks fine" and "actually optimized" is where most e-commerce accounts leave money.
Feeds decay. Products change, seasons shift, competitors rewrite their titles, Google updates its requirements. A feed optimized once and left alone drifts back toward mediocrity — usually without anyone noticing, because nothing visibly "breaks."
The stores that win Shopping treat the feed like a living asset: reviewing search term reports to find vocabulary gaps in titles, testing title structures on top sellers, updating custom labels as margins shift, and auditing attribute completeness quarterly. None of it is glamorous work — which is exactly why it's a durable advantage. Most competitors won't do it.
A practical objection to feed work is that product data lives in the shop system, where changing titles affects the website itself. Supplemental feeds solve this: they let you overwrite or enrich specific attributes — titles, custom labels, product types — for advertising purposes only, without touching your site. Combined with Merchant Center feed rules, this means feed optimization can run as its own workstream, iterating weekly, while your product catalog stays untouched. There is no infrastructure excuse for a stagnant feed.
If you're not sure where your feed stands, that's a measurable question, not a matter of opinion. A professional Google Ads audit reviews your feed quality alongside your campaign structure and tracking — or a free Google Ads audit gives you a first read on what's holding your Shopping performance back.
Because in Shopping and Performance Max, the feed determines which auctions you can enter at all. Bids only influence how competitively you show up in the auctions you're eligible for — the feed decides eligibility itself. A perfect bidding strategy applied to a weak feed optimizes a small, poorly matched slice of your potential traffic, and raising bids on poorly described products just means paying more for the same wrong auctions.
Product titles, starting with your best sellers. Rewrite them to front-load brand, product type, and the attributes customers actually search — gender, color, size, material — instead of internal model codes. It's the highest-leverage change because it improves both the volume and the relevance of your impressions simultaneously, and the search terms report for your top products tells you exactly which vocabulary is missing.
Wherever a GTIN exists, submit it. It connects your product to Google's catalog knowledge and consistently improves performance; missing or wrong GTINs can quietly limit your visibility. For genuinely custom or own-brand products without barcodes, set the identifier-exists flag correctly instead of leaving the field blank — and never invent GTIN values, which is a fast route to disapprovals.
Availability and price should sync as close to real time as your platform allows — daily at minimum, and tighter during sales periods when prices move fastest. Content optimization is ongoing work: review search terms monthly for vocabulary gaps, clear Merchant Center diagnostics monthly, and audit attribute completeness at least quarterly. A feed is never "done" as long as your catalog and competitors keep changing.
Yes. Feed quality feeds into relevance, and weaker relevance means lower click-through rates and worse auction positioning — which you compensate for by paying more per click. Conversely, a well-optimized feed earns more relevant impressions and better engagement, effectively lowering the price you pay for the same customer.
Custom labels (custom_label_0–4) don't affect matching at all — Google ignores them for relevance. They exist so you can segment your catalog inside campaigns: by margin band, price bracket, performance tier, seasonality, or stock depth. That makes them the bridge between feed and strategy — they're what let you give high-margin products their own budget and target instead of averaging the whole catalog into one blended number.
Yes — supplemental feeds and Merchant Center feed rules let you overwrite or enrich attributes like titles, product types, and custom labels for advertising purposes only, without touching your shop system. Feed optimization can run as its own iterative workstream while your site stays untouched, so infrastructure is never a valid reason for a stagnant feed.
Three quick checks: open Merchant Center's item-level diagnostics and total up the revenue of disapproved and limited items; scan the search terms report for both irrelevant matches (over-claiming) and missing high-intent queries you know exist (under-claiming); and compare a sample of your titles against how customers actually search your category. If any of these turn up meaningful gaps, the feed is capping you — and a structured review will quantify by how much.
Online marketing and PPC specialist focused on Google Ads and advertising strategy — the kind that builds not just clicks, but brands. With 10+ years in digital marketing and e-commerce, Tom leads on strategy and measurement, turning strategic scaling and zero-click trends into measurable business results.
in
Performance marketing specialist with deep hands-on Google Ads and e-commerce experience. László leads on campaign execution and growth, building and scaling accounts for e-commerce brands and small businesses — the same senior specialist on your account from day one, not a junior and a dashboard.
in
By
·
1 minute read