To structure Shopping campaigns, we consider how products may be naturally organized, how bids and budgets can be managed, and how easily performance can be analyzed as campaign data accumulates. For eCommerce advertisers with hundreds or thousands of products, a well-designed structure can make a substantial difference in how manageable the account becomes.
Unlike Search campaigns, Shopping campaigns don’t primarily rely on advertiser-selected keywords to determine when ads are eligible to appear. Instead, Google uses information contained in the advertiser’s product feed, along with campaign settings and other signals, to match products with relevant searches.
This makes the relationship between the product feed and campaign structure particularly important. The feed tells Google what the products are, while the campaign structure gives the advertiser a framework for organizing, bidding, analyzing, and managing those products.
Individual Shopping campaigns are structured internally on two important levels: product groups, and ad groups. Both can play a role in organizing products, managing bids, evaluating performance, and improving how search queries match to the appropriate products in the feed.
Product Groups
Product groups provide a mechanism for grouping similar products within a Shopping campaign. Instead of treating every SKU as an entirely independent advertising unit, products can be subdivided using attributes contained in the Merchant Center product feed.
Depending on the available product data, products can be grouped using attributes such as brand, product type, Google product category, item ID, condition, channel, and custom labels.
For example, an online retailer selling several brands might initially subdivide its inventory by brand. Individual brands could then be divided further by product type or another useful feed attribute.
A product group can contain a single SKU, a group of closely related products, product variants, an entire product category, or a much larger portion of the feed. The appropriate level of segmentation depends on the size and characteristics of the catalog, available advertising budget, and the amount of control required.
This becomes especially valuable with sizeable product feeds. Managing thousands of products individually would be unnecessarily cumbersome in many accounts. Product groups allow similar products to be managed together while preserving the ability to isolate strategically important products when necessary.
Product Groups and Bidding
Product groups also provide an important level of bidding control in Standard Shopping campaigns.
Products with similar economics can often be grouped together so that bids reflect their relative value to the advertiser. A retailer might, for example, want different bidding treatment for premium products, high-margin products, clearance inventory, best sellers, or products with very different average order values.
This is one reason feed design and campaign design should not be treated as completely separate activities. Useful attributes in the feed give the advertiser more options for creating meaningful product groups within the campaign.
As performance data accumulates, product groups can also be subdivided further. A group that initially contains dozens of similar products may eventually reveal several high-performing SKUs that warrant separate treatment.
Conversely, creating excessive segmentation before enough traffic exists can make a campaign unnecessarily complicated. The objective is to create enough structure to support useful control and analysis without dividing the available traffic into so many small segments that meaningful performance patterns become difficult to identify.
Ad Groups in Shopping Campaigns
Ad groups provide another organizational and control layer within Standard Shopping campaigns.
They can be particularly useful because negative keywords can be assigned at the ad group level. This gives advertisers an additional mechanism for influencing which portions of the product catalog are eligible for particular types of searches.
Suppose an eCommerce advertiser sells several closely related product families. Google may consider products from more than one family relevant to a particular search, even though one group represents a substantially better match to the shopper’s intent.
Separating those product families into appropriate ad groups can make it possible to use negative keywords to reject an undesirable match in one ad group while allowing the search to reach the more appropriate products in another.
This is conceptually similar to the role negative keywords play in Search campaigns, although Shopping campaigns use product data rather than conventional positive keyword targeting to establish product eligibility.
Using Negative Keywords to Improve Shopping Query Matching
Negative keywords can be especially valuable in Shopping because advertisers do not have the same direct keyword controls available in conventional Search campaigns.
Google determines which products are potentially relevant to a query from the product data and other signals. That automation can produce excellent matches, but it can also produce searches that are only loosely aligned with the product being advertised.
Negative keywords give the advertiser a way to reject some of those unwanted matches.
When ad groups are designed around meaningful product distinctions, negative keywords can become more precise. Instead of excluding a search from the entire campaign, an advertiser may be able to prevent the query from reaching the wrong product group while retaining eligibility for products that represent a better match.
This approach can help reduce wasted clicks and improve the relationship between shopper intent and the products displayed in Shopping results.
Using Product Types to Structure Shopping Campaigns
The product type attribute is one of the most useful advertiser-controlled fields in an eCommerce product feed.
Unlike Google’s predefined product category taxonomy, product type allows the merchant to organize products according to its own catalog structure and business logic. Product types can use multiple hierarchical levels, allowing products to move from broad classifications into increasingly specific subcategories.
A simplified hierarchy might look something like:
Electronics > Audio > Headphones > Wireless Headphones > Premium Wireless Headphones
This hierarchy can provide useful information for campaign segmentation, feed analysis, reporting, and optimization. A well-designed product type structure makes it easier to analyze groups of related SKUs and can provide another useful attribute for subdividing inventory within Shopping campaigns.
Product type is also one of several feed attributes that can help Google understand the nature of an individual SKU. For more detail about titles, descriptions, categories, product types, and other product data, see our guide to product feed optimization.
Using Custom Labels for Business-Oriented Structure
Not every useful way of organizing products corresponds to a conventional product category. This is where custom labels can become particularly valuable.
An advertiser might want to distinguish products according to profitability, seasonality, inventory levels, sales volume, promotional status, or some other characteristic that matters to the business but is not represented by Google’s standard product attributes.
For example, products could be labeled internally as high-margin, core inventory, seasonal, clearance, best seller, or new product. Those labels can then support campaign and product group decisions.
This allows Shopping campaign architecture to reflect the advertiser’s business priorities rather than merely mirroring the structure of the website or Google’s product taxonomy.
Structuring at the Shopping Campaign Level
There are many possible approaches when segmenting an eCommerce feed into individual Google Shopping campaigns.
We begin by examining the feed and understanding the relationships among the products. Product categories, brands, prices, margins, search behavior, historical performance, and business priorities can all influence the appropriate campaign architecture.
From there, we determine which products can reasonably serve together and which products warrant separation into different campaigns.
Campaign-level separation can be useful when different portions of the catalog require independent budgets, substantially different bidding strategies, different geographic targeting, or other forms of independent control.
Individual products can then be assigned into product groups based on product characteristics, pricing, feed attributes, and other relevant distinctions.
Where appropriate, product groups can be further segregated using ad groups in order to take advantage of negative keywords and make the campaign more selective. With negative keywords customized at the appropriate level, search queries can be directed away from poor product matches while preserving stronger matches elsewhere in the campaign.
Shopping Campaign Structure Should Reflect Available Data
More segmentation is not automatically better.
A very large eCommerce account with substantial traffic and conversion volume may support a highly granular structure. A smaller advertiser selling the same number of products but generating much less traffic may need a simpler architecture so that useful performance data can accumulate.
Campaign structure therefore needs to balance control against data density.
Too little segmentation can hide meaningful differences between products. Too much segmentation can create a complicated account containing many product groups or campaigns with insufficient traffic to make reliable optimization decisions.
The right structure usually becomes clearer as actual campaign data accumulates. Strong performers can be isolated, weak products can be identified, and budgets can increasingly be directed toward areas of the catalog producing the greatest business value.
Shopping Structure and Performance Max
Standard Shopping and Performance Max use different campaign architectures, but the underlying quality and organization of the product feed remains important to both.
Performance Max uses Asset Groups and Listing Groups rather than the traditional Shopping campaign structure described above. It also gives Google’s automation considerably more control over targeting and ad delivery.
Nevertheless, the work involved in organizing products, improving feed attributes, understanding SKU-level performance, and establishing meaningful product categories can carry over directly into Performance Max campaigns.
For large eCommerce catalogs in particular, understanding the relationship between feed structure and campaign structure becomes increasingly important regardless of which Shopping-oriented campaign type is being used.
Campaign Structure Evolves With Performance
To structure shopping campaigns, the right approach provides a strong foundation; but it doesn’t have to remain static. It can improve.
Once the campaign has launched and accumulated meaningful data, performance can reveal distinctions that were not obvious during initial campaign development. Certain product groups may consistently outperform others. Individual SKUs may account for a disproportionate share of revenue. Some categories may require different bidding or budget treatment.
That information can be used to refine the campaign structure over time.
This is why campaign architecture and ongoing PPC campaign management are closely connected. Structure provides the framework for collecting useful data; performance data then provides the information needed to improve that structure.
Related Shopping Campaign Resources
For more information about building and managing eCommerce PPC campaigns, see our resources covering Google Shopping campaigns, product feeds, product feed optimization, Performance Max campaigns, and PPC campaign management.
