AI Max is now available in beta for Standard Shopping campaigns, introducing artificial intelligence capabilities that, until now, we primarily associated with Search and Performance Max. Standard Shopping campaigns are becoming less common among our clients, partly because Google encouraged e-commerce businesses to rely mainly on PMax, promising greater profitability.
Let's explore what this new AI Max layer brings to Shopping campaigns. Just like AI Max for Search campaigns, enabling it is entirely optional and can be configured at the campaign level.
Until now, Shopping ads used the product titles defined in the Merchant Center feed. With AI Max, Google can personalize those titles based on the user's search intent and search history.
For example, imagine a hair dryer with the title "30W Hair Dryer." If someone searches for "cordless travel hair dryer," Google could generate a title highlighting those features, provided it can find supporting information on the website. This is another reason why making your website AI-readable is so important.
This is arguably the most disruptive new feature. If you enable final URL expansion, a Standard Shopping campaign can start displaying text ads alongside traditional product ads.
Google can identify other pages on your website that it considers relevant, such as category pages or new arrivals sections, to capture searches that don't necessarily lead to a specific product page.
Unfortunately, page feed controls aren't as advanced as those available in AI Max for Search. Therefore, if we enable URL expansion, we need to carefully monitor which pages receive traffic and how they perform.
This means explicitly excluding pages that don't serve a commercial purpose, such as contact pages, blog posts, or any other URLs we don't want to promote.
Additionally, this feature requires using a Target ROAS bidding strategy and enabling text customization.
AI Max for Shopping also introduces brand exclusions. These allow advertisers to exclude certain brands from text ads while, where applicable, still allowing their products to appear in Shopping ads.
This is particularly useful because many accounts already have dedicated Search campaigns targeting branded queries.
If we allow our Shopping campaigns to capture that same traffic through text ads, we could end up mixing results and attributing conversions to AI Max that were previously generated by other campaigns.
The more complete your Merchant Center data is, the more information Google has to personalize your ads. Attributes such as materials, dimensions, technical specifications, and product compatibility can be particularly valuable.
However, once we enable AI Max, optimizing the product feed alone will no longer be enough.
✅ Search terms and additional traffic identified as AI Max: Add negative keywords using Dolnai's N-Gram analysis.
✅ Automatically selected landing pages: Exclude pages that perform poorly or aren't designed to drive conversions.
✅ Google-generated product titles and assets.
✅ Brand and URL exclusions.
✅ ROAS: Monitor not only campaign-level ROAS but also the overall account ROAS.
Google claims that AI Max for Shopping delivers, on average, 5% more conversions or conversion value, but that doesn't mean it should automatically be enabled across every account.
Results may vary significantly from one account to another.
Google Ads' more traditional campaign types—Search and Shopping—now both benefit from AI Max, offering a level of configuration and targeting that can be particularly appealing to experienced advertisers.
Why concentrate our entire strategy on Performance Max when we can combine AI-powered Search and Shopping campaigns while maintaining greater control over certain decisions?
Although PMax remains the campaign type that enables advertisers to work across Google's different channels within a single campaign, it may make sense to give Shopping a more prominent role and use PMax as a secondary campaign or a complementary strategy.
That said, these AI Max technologies are still too new to reach a definitive verdict.
.svg)
