Product as a data object in Google & Microsoft Ads

Most performance-driven campaigns in Google or Microsoft Ads today run with maximum automation. However, many marketers face a dilemma:

…why do some products “take off” while others “sleep,” even with the same budget and settings?

The answer lies in how these platforms see and process a product as a data object, not simply as a “type of goods” or a nice image.

Advertising system algorithms do not see products the way humans do—they do not evaluate creativity or emotion. They see structured data coming from the product feed in Google or Microsoft Merchant Center. This feed determines whether a product can be shown at all, how it is compared to competitors, and how bidding responds.

For a product to be fully performance-ready, it must have the correct identification attributes:

Source: Google Support

These identifiers help the algorithm match the product to relevant demand and compare it correctly within the market. Products with valid GTINs often achieve better visibility and higher relevance because the system can reliably place them into auctions.

Source: Google Support

If a product does not have the required identifiers available (GTIN, or alternatively MPN and brand in a sufficient combination), this must be explicitly communicated to Google using the attribute identifier_exists = no (or false)

This confirms that identifiers genuinely do not exist for the product. Inventing or using invalid identifiers (e.g., a fake GTIN) is inappropriate and may lead to product disapproval or reduced visibility.

For the algorithm, the product title acts as a kind of meta-navigation signal. It should be:

Inconsistent or “SEO-heavy” titles can reduce search relevance and worsen CTR. On the other hand, a clean and clear title gives the system a better match with user queries.

Source: LinkedIn GoDataFeed

Algorithms do not just compare prices—they use them:

Source: Google Support

Availability (“in stock” vs. “out of stock”) also directly affects product visibility. Out-of-stock products are often suppressed or not shown at all. That’s why it is critical for the feed to contain up-to-date availability and pricing data.

Source: Google Support

Variants such as colors, sizes, or bundles are not just aesthetic. The algorithm treats them as separate SKUs (Stock Keeping Units), each with its own performance. If variants of the same product have inconsistent or contradictory data in the feed, the system may:

Properly separated and clearly defined variants help the algorithm learn faster and optimize more effectively.

Although a product feed may appear to be a static dataset, algorithms also incorporate performance signals such as:

Products with strong historical performance are then prioritized in budget allocation and tend to receive more impressions under the same conditions than lower-performing products.

Technical marketers often overlook this issue, yet discrepancies between feed data and on-site reality are among the main reasons systems lower a product’s performance score. Most commonly, differences occur in:

Such inconsistencies can lead to Merchant Center penalties, reduced impression share, and lower relevance of product listings.

Like Google, Microsoft Advertising also relies on the product feed as its primary data source. The difference is that it uses its own algorithm to evaluate products, even though it can directly import feeds from Google Merchant Center.

This approach simplifies data synchronization across platforms and ensures that key information — product titles, prices, availability, and inventory status — remains consistent in both systems.

Feed quality is just as critical here. Inconsistent or outdated data can prevent Microsoft Advertising from delivering products correctly, reduce visibility, or negatively impact campaign performance. The same rule applies as with Google Ads: the more accurate and consistent the data, the better the algorithm can optimize impressions, bidding, and product relevance.

Campaign performance is not just about budgets, settings, or bidding strategies. At its core, it depends on how well your products “speak the language” of the algorithms that decide initial visibility, relevance to specific queries, and the efficiency of automated strategies.

When working with products, it is therefore essential not only to have a properly structured feed but also to ensure its consistency and freshness. High-quality data enables systems to correctly compare products, place them into auctions, and optimize their performance.

Since feed quality is the foundation of success, it’s worth using tools that help with this process. BlueWinston offers the ability to create or modify a product feed directly from your Google or Microsoft Merchant account, significantly simplifying synchronization and ensuring consistent data across platforms. We described this functionality in detail in the article “Don’t Have a Product XML in BlueWinston? No Longer a Problem.”

Only when you understand products as data objects with signals do you gain a solid foundation for:

This approach allows the system to work with your products as effectively as possible and maximize their performance with the same budget and settings.