Zombies Have Taken Over Google

Google Performance Max (PMAX) campaigns have become a central driver in maximizing e-commerce revenue for brands. This automated, algorithm-driven approach has proven highly effective for traditional shopping campaigns. However, as with many Google products, there are limitations in terms of an advertiser's ability to control spend and performance at a granular level. One significant challenge within the "Google Black Box" is that low-volume products are often overshadowed by the best-performing SKUs, becoming part of a "Product Zombie Nation," where revenue plateaus. Let’s explore how this works, and what you can do to avoid the Zombie Apocalypse in your PMAX campaigns.

PMAX Best Practices

PMAX allows product feeds to reach every media type Google offers, all through a small number of well-structured asset groups. It enables broader audience reach and utilizes a variety of ad formats, optimizing for the best-performing combinations using AI.

Although it’s not explicitly stated in official Google documents, Google tends to favor PMAX over traditional "standard" shopping campaigns. If you try to run both in tandem, the traditional shopping campaign will quickly lose traffic. A PMAX campaign will almost always "bulldoze" a standard shopping campaign, making it an inefficient strategy to run them together.

Thus, PMAX has become essential for e-commerce businesses to remain competitive in today’s ad-driven ecosystem. One key takeaway from PMAX is that product category consolidation—using a broad, "catch-all" campaign structure—works better than segmenting products or SKUs in great detail. This often goes against the instincts of search marketers, as it limits control, and is the opposite of how traditional search marketing and shopping campaigns are structured. Typically, those campaigns benefit from tight segmentation, with product-based ad groups and tailored ad copy designed to respond to individual searches with closely matched keywords.

So when search marketers were first told to build broad, multi-category PMAX campaigns with as many SKUs and creative assets as possible, trusting the algorithm to match them to prospective intent, we were skeptical. However, after seeing the performance gains and predictable high ROAS (Return on Ad Spend) revenue, we began to embrace the "cast a wide net" approach.

A New Day Has Dawned

With recent algorithm changes and Google continually increasing CPCs (cost-per-click), the once-reliable "wide net" strategy has been yielding diminishing returns for many businesses. The challenge we’re seeing with this approach over time is that as a PMAX campaign runs longer, it tends to push high sales volume products to the forefront. While this ensures that your best-performing products always get traffic, it reduces the chances that other products in the campaign will receive impressions—especially when there are budget constraints.

Products that don’t generate sales, despite being important to the client or offering high margins, are left in the shadows. These products effectively become "dead," receiving no impressions and having little to no chance of generating revenue. In essence, they might as well not exist within the campaign.

Time to Feed the Zombies

To overcome this challenge, we shifted our strategy. We tested a new approach by moving from a single PMAX campaign to a complementary, two-tiered strategy. Our goal was to "resurrect" our dead products by giving them their own dedicated campaign. This campaign for low-traffic SKUs is affectionately referred to as a "Zombie" campaign. It includes only products that have seen little to no traffic or sales in the past year.

To set up this "Zombie" campaign, we first downloaded the performance data from our legacy PMAX campaigns, filtered by product. We then removed any products that had made sales and used the remaining products to create the Zombie campaign structure. The third and most crucial step was to exclude these dead products from the original PMAX campaign, ensuring both campaigns could run without duplication.

This approach has proven successful with our current clients. For example, one client launched the Zombie PMAX campaign in mid-October and ran it through the end of the year. This Zombie PMAX campaign drove 980 sales, generated $133k in revenue, and maintained an overall ROAS of 618%. This performance exceeded expectations, surpassing the ROAS of our legacy PMAX campaign, and generated 22% of total sales and revenue. Redirecting the shopping budget into the Zombie PMAX campaign resulted in an additional $21k in revenue—revenue we wouldn’t have generated had the budget been allocated to the legacy PMAX campaign (which had an average ROAS of 525%).

Summary:

The new two-tier campaign approach has proven successful for Carbon clients:

And just like that, the "dead" products came back to life—proving that even in e-commerce, sometimes a little zombie resurrection is all it takes to spark new growth and avoid being The Walking Dead to the board room to report on quarterly sales results.