Product Feed Fundamentals: What Google Actually Needs

In the ecosystem of Google Ads for UK eCommerce, your product feed is the singular source of truth. It is not merely a technical file; it is the fundamental vocabulary that allows Google's AI to understand what you sell, who should see it, and why your SKU is the best match for a specific search intent.

A common misconception among retailers is that as long as the Merchant Center light is green, the feed is "optimised." This is dangerously incorrect. A "green light" only means you have met the bare minimum technical requirements to enter the auction. To win the auction - and to do so at a profitable ROAS - you must engineer a feed that provides high-fidelity signals to Google's smart bidding algorithms.

Google requires 11 attributes as a minimum. Additional 15+ recommended attributes unlock advanced segmentation, reporting, and smart-bidding signal quality. This article walks through every attribute (what it does, why Google needs it, and how to populate it correctly for UK eCommerce), then compares the 5 feed source options - from Shopify's native sync to the Content API - ranked by control and engineering effort required.

How Merchant Center Reads Your Data

Google Merchant Center (GMC) acts as the bridge between your website and your advertising account. It is essentially a parsing engine. When you provide a feed, whether via XML, TSV, or API, Google runs a three-stage validation process:

1. 1
   The Parsing Phase

   The crawler identifies your data format. It looks for header keys (like "price" or "id") and maps them to its internal database schema. If your headers are non-standard, GMC attempts to "guess" the intent, which is where many errors begin.

2. 2
   The Validation Phase

   The system checks for policy and formatting compliance. For example, it verifies that the price matches the currency of the target country and that the image_link leads to a valid, reachable image file without promotional watermarks.

3. 3
   The Synchronization Phase

   Valid products are indexed into the Shopping Graph. Every hour, Google Ads fetches this indexed data to populate Shopping Ads, Performance Max campaigns, and the organic Shopping tab.

In the UK, the freshness of this sync is critical. Due to volatile exchange rates and high competition, price mismatches are the leading cause of account-level warnings. Google expects the data in your feed to be a perfect reflection of what the user sees on your landing page at the moment of the click.

The 11 Required Attributes

These are the "non-negotiables." If even one is missing or incorrectly formatted, the product will be disapproved.

Recommended Performance Attributes

If the required attributes are the skeleton, these recommended fields are the muscle. They don't just "help"; they are the primary drivers of campaign segmentation and ROAS.

The product_type attribute is your best friend for campaign structure. While google_product_category is for Google, product_type is for you. Use it to reflect your site's navigation (e.g., Living Room > Seating > Velvet Sofas), allowing you to subdivide Performance Max campaigns with surgical precision.

Furthermore, custom_labels allow you to pass non-product data into your bidding environment. We frequently use these to label products as "High Margin" vs "Low Margin," enabling us to set different tROAS targets within the same account.

item_group_id deserves particular attention for any retailer selling variants - size runs, colourways, or bundle configurations. Grouping variants under a shared item_group_id allows Google to present them as a single "swatch" unit in the ad, letting the shopper pick a colour or size directly within the ad surface rather than clicking through and bouncing when their preferred variant is unavailable. Retailers who fail to set this attribute correctly often see each variant compete against its own siblings in the auction, artificially inflating CPCs for products that should never have been bidding against each other in the first place.

shipping_weight and the wider shipping attribute set (shipping, shipping_label) feed directly into Google's landed-cost calculations, which increasingly determine whether your product surfaces for cost-conscious searches. As free and fast delivery become table-stakes expectations in UK eCommerce, an incomplete shipping attribute set means Google falls back to your Merchant Center account-level shipping settings, which are rarely granular enough to reflect genuine cost variance between a lightweight accessory and a bulky item requiring a two-man delivery service.

Finally, availability_date is the most under-used attribute in UK preorder-heavy categories such as consumer electronics, gaming and seasonal homeware. Retailers who list a preorder product without this attribute risk Google treating the stock status as ambiguous, which suppresses the listing during exactly the high-intent window (pre-launch hype) when it should be performing best. Setting an accurate availability_date lets Google continue serving the ad with full confidence right up to the fulfilment date.

Feed Quality Signals & Data Hygiene

Google does not evaluate your feed as a static checklist. It evaluates it continuously, comparing your data against three things: your own website (via the Google crawler), your historical performance, and the feeds of competing retailers selling comparable products. This comparison produces what we internally call "quality signals" - proxies the algorithm uses to decide how much to trust your data, and by extension, how much impression volume and bid headroom to grant you.

The first signal is consistency. Every time Googlebot crawls your landing page, it checks the price, availability and title against what your feed submitted. A feed that is accurate 99% of the time but drifts on flash-sale days will accumulate a "low trust" flag that is invisible in the Merchant Center diagnostics UI but visible in reduced impression share. We have seen accounts recover 15-20% lost impression share within a fortnight of eliminating recurring price drift, simply by tightening the sync cadence between the cart platform and the feed source.

The second signal is completeness. Google explicitly rewards feeds that populate optional fields relevant to the product category. Apparel listings without size, colour and age_group are eligible for fewer of the visual, filterable Shopping surfaces (such as the "Browse" grid experience) than fully attributed competitors. This is not a disapproval - the item still runs - but it is a silent tax on reach that most retailers never notice because Merchant Center will not warn you about a missing optional field the way it warns you about a missing required one.

The third signal is competitiveness, surfaced directly in the "Price competitiveness" and "Benchmark" reports inside Merchant Center. These reports compare your price and CTR against the median for identical GTIN-matched products across the wider Shopping Graph. Retailers who are structurally more expensive than the market rarely win the auction on Shopping placements regardless of how well-optimised their attributes are, because the buy-side signal (predicted CTR at that price point) is baked into the ranking. If your margin structure means you cannot compete on price, the correct response is not to abandon Shopping, but to lean harder on the attributes that do not depend on price - richer imagery, tighter titles, and custom labels that let you bid more aggressively only on the SKUs where you are genuinely competitive.

A fourth, less discussed signal is freshness decay. Feeds that are only refreshed once every 24-48 hours accumulate a growing gap between the advertised state and the true state of stock and pricing. For retailers running frequent promotions, this gap is where disapprovals and mismatched-price warnings cluster. The fix is rarely a full platform migration; it is usually tightening the fetch schedule (moving from a daily manual upload to an automated twice-daily or hourly scheduled fetch) so that the delta between reality and feed shrinks to a window Google is comfortable serving against.

UK Market Specifics (VAT & Shipping)

Operating in the UK market introduces several unique requirements that, if ignored, lead to account suspension.

Character Limits & Formatting Rules

Beyond simply populating the correct attributes, Google enforces strict formatting rules on almost every field. Getting these wrong does not always trigger an outright disapproval - often the item is silently truncated or auto-corrected, which quietly damages performance without ever appearing as an "error" in your diagnostics dashboard.

The most common formatting mistake we see in UK feeds is unescaped HTML inside the description field. Content management systems that let merchandisers write rich-text product copy will often export raw <br>, <p> or <strong> tags straight into the feed. Google's parser will either strip these (leaving garbled, run-on text) or, in stricter cases, disapprove the item entirely for "promotional text" violations if the tags wrap phrases like "free shipping" or "buy now." The fix is a simple strip-tags transformation at the point the feed is generated - most feed management apps (Channable, Feedonomics, DataFeedWatch) include this as a built-in rule, but native platform exports frequently do not.

Titles present a subtler formatting trap: Google explicitly disallows promotional phrases such as "Best Seller," "% off," or excessive capitalisation ("SALE NOW ON") in the title field. These will not always trigger disapproval, but they are frequently down-weighted in the ranking algorithm because Google treats them as low-quality signal rather than descriptive product information. Reserve promotional messaging for the sale_price mechanism, which surfaces a genuine strikethrough badge in the ad unit rather than a keyword you have manually inserted into the title.

Advanced Feed Engineering

Once you have established the basics, the next level is feed engineering. This is the practice of manipulating your data specifically to improve your ad performance, rather than just describing the product.

Semantic Title Injection

Most Shopify stores sync titles exactly as they appear on the site. However, searchers on Google use different intent signals. If your product is named "The Serenity Ring," it carries zero SEO weight. Feed engineering involves appending high-intent keywords: "18k Gold Diamond Engagement Ring - The Serenity - UK Size M." By placing the most relevant keywords in the first 70 characters, you dramatically increase your click-through rate.

Price-Competitive Bidding Labels

Google Merchant Center provides a "Price Competitiveness" report. By pulling this data into a supplemental feed via a tool like Feedonomics or Channable, you can create a custom label for "Cheaper than Competitors." In our eCommerce strategy, we often bid 50% more aggressively on products where the retailer is the cheapest in the market, as these products naturally have a much higher conversion rate.

Feed Source Options Ranked

There is no single "correct" way to get product data into Merchant Center. The right choice depends on catalogue size, how frequently your pricing and stock levels change, and how much engineering resource you have available to maintain the pipeline. Below we rank the five most common feed sources by the control they give you over the data and the ongoing effort required to keep that data accurate.

6.1 Shopify Merchant Center Native Sync

Shopify's built-in Google & YouTube channel app pushes your catalogue directly into Merchant Center with minimal setup - genuinely a few clicks for a store already using standard Shopify fields. The trade-off is control: you cannot easily override individual attributes without app-level workarounds, and the mapping logic (particularly for google_product_category and gtin) is a black box. For a starter store with a simple, well-tagged catalogue under a few hundred SKUs, this is the sensible default. For anything larger or with pricing complexity (multi-currency, wholesale tiers, frequent sales), native sync becomes a ceiling rather than a foundation.

6.2 WooCommerce Feed Apps

WooCommerce has no native Merchant Center integration, so retailers rely on dedicated feed plugins (such as Product Feed PRO or CTX Feed) that generate and host an XML/CSV file from your WordPress product data. These apps offer good mid-tier control - rule-based attribute mapping, category overrides, and scheduled regeneration - without needing developer resource. The main risk is server load: on shared hosting, large catalogues regenerating feeds hourly can slow the storefront itself, so feed generation should be offloaded to a cron job during low-traffic hours.

6.3 Magento Feed Extensions

Magento (Adobe Commerce) retailers typically use marketplace extensions or dedicated feed management platforms such as DataFeedWatch or Channable, layered on top of Magento's native product export. These tools give the deepest rule-based transformation capability of any non-developer option: conditional logic, bulk find-and-replace across titles, and automated category mapping against Google's taxonomy. This is the natural home for mid-to-large UK retailers (roughly 1,000-20,000 SKUs) who need serious feed engineering without building custom infrastructure.

6.4 Google Sheets Feed

For very small catalogues, a manually maintained Google Sheet registered directly as a feed source in Merchant Center remains a perfectly legitimate option. It offers absolute control - every cell is exactly what you typed - at the cost of being entirely manual. Sheets feeds do not scale past a hundred or so SKUs before the maintenance burden and error rate (a mistyped price, a stale availability flag) outweigh the simplicity. We generally recommend Sheets only as a stopgap while a proper feed pipeline is being built, or for boutique retailers with genuinely static catalogues.

6.5 Content API for Shopping

The Content API is the developer route: a direct programmatic connection between your product database (or ERP) and Merchant Center, bypassing scheduled file fetches entirely. Changes to price, stock, or promotional status can push through in near real time, which matters enormously for high-velocity inventory such as flash sales, flight/travel pricing, or limited-run drops. The cost is engineering time - this is not a marketing-team configuration, it is a development project requiring OAuth setup, error handling, and ongoing maintenance as Google periodically updates the API schema. Enterprise retailers running 10,000+ SKUs with volatile pricing are the clearest use case; smaller retailers rarely see enough incremental ROAS to justify the build cost.

Feed Update Frequency

How often your feed refreshes is, on its own, a ranking-relevant signal. Google explicitly favours retailers whose data is demonstrably fresh, because a fresher feed reduces the risk of serving an ad for an item that is actually out of stock - a poor user experience Google is highly motivated to avoid.

Real-time, via the Content API, is the gold standard for stores with fast-changing inventory: flash sales, limited drops, or marketplaces reselling third-party stock where availability can change within minutes. Daily scheduled fetch - Merchant Center pulling your hosted feed file once or twice every 24 hours - is the right cadence for the vast majority of standard eCommerce catalogues, where stock and pricing change at a manageable pace. On-demand manual upload should be reserved for very small catalogues, generally under 50 SKUs, where a human can realistically keep the file accurate without automation.

The practical risk of infrequent updates is not just missed sales opportunities - it is active policy risk. If your scheduled fetch runs once every 48 hours and a product sells out six hours after the last fetch, Google continues serving ads against a phantom listing for up to two days. A pattern of these "availability" mismatches, even if each individual instance is brief, accumulates into an account-level trust penalty that suppresses impression volume across your entire catalogue, not just the affected SKUs. For any UK retailer running regular promotions or dealing with genuinely finite stock, moving from daily to twice-daily (or hourly) scheduled fetches is one of the highest-leverage, lowest-cost changes available - it typically requires nothing more than adjusting a setting in your feed app or Merchant Center's scheduled fetch configuration.

Maintenance & Common Mistakes

These are the 6 most common errors that drain budget in UK Merchant Center accounts. What makes them dangerous is that most do not produce an outright disapproval - they quietly suppress reach or inflate cost-per-click while the account otherwise looks healthy in the diagnostics dashboard.

• Generic Categorization: Mapping everything to Home & Garden instead of the specific sub-category like Home & Garden > Kitchen > Cookware.
• Unescaped HTML: Leaving <br> tags in the description. Google often disapproves items with raw code.
• Broken Image Links: Changing image structure without updating the feed, leading to 404s.
• Price Mismatches: The feed sends £100 but the landing page shows £120 (VAT issues).
• ID Instability: Changing SKUs frequently, which kills historical performance data.
• Missing Sale Dates: Not using sale_price_effective_date, causing sale prices to linger.

Frequently Asked Questions