Brunner Recommends Rigorous Testing Before Going with ‘PMax’
Google launched Performance Max in late 2021 with lots of fanfare about the benefits of AI-powered online advertising.
More than two years later, reviews are mixed. At Brunner, we manage paid search engine marketing (SEM) for clients of all types with significant monthly spends. We’ve seen strong performance from PMax, but poor performance, too — and without data insights to understand how and why.
We recommend that advertisers exercise caution using PMax, and we’re not the only ones. Read on for lessons learned from our team’s experience as a PMax early adopter and evolution into a power user, and from our perspective as a Google Premier Partner.
How Google Performance Max works
Google Performance Max is a goal-based ad format that can run across Google-owned channels, powered by artificial intelligence and machine learning. Advertisers, or agencies like Brunner that manage Google Ads accounts for clients, upload multiple asset types, including text, images, videos, and logos, in one interface. PMax can then serve ads created from these assets across multiple Google channels, including YouTube, Display, Search, Discover, Gmail and Maps.
Performance Max campaigns run automatically based on an advertiser’s goals. From those inputs, Google chooses which users to target, when to show an ad, what text to use and on which channels. Google controls how each ad is assembled, how it is bid on in each auction — and how much of the overall budget is spent on each channel.
All that automation adds up to enormous control for Google. It also means far less insight for advertisers and agencies, especially when results fluctuate. We’ve learned that testing and vigilance is essential.
“Thorough testing and informed decision-making are paramount in navigating the complex landscape of online advertising with Google’s PMax.”
PMax for Lead Gen: Excessive Bot Activity and Poor-Quality Leads
As an early adopter of Google’s PMax, we encountered significant challenges at first. We noticed immediately that using PMax to generate leads led to excessive bot activity and poor-quality leads. At the surface level, PMax seems like a no-brainer for producing a high volume of efficiently generated leads. But when you look a little closer, you’ll find a different story.
For example, one test campaign generated a high number of site lead form fills and click-to-calls — at a desirable cost per lead. We used CRM software and call tracking services to verify the leads. We quickly found:
- less than 1% of the “leads” generated from PMax were real leads.
- most click-to-call button clicks did not even result in a call.
- site leads from fill submissions were filled with fake information, a telltale sign of bot activity.
Worse still, the sudden influx of poor-quality leads put a strain on our client’s call center and sales team as they tried to make sense of the poor-quality leads.
Lesson learned: Use Call Tracking, CRM Integration with PMax
One crucial lesson from testing PMax lead gen campaigns is the importance of having call tracking and/or CRM integration in place, so you can understand lead quality.
In theory, connecting Google Ads to a CRM system to flag quality leads should fuel Google’s algorithm to drive more quality leads through PMax automation. Unfortunately, we found that PMax could not even generate enough quality leads to inform algorithmic optimization. That resulted in extremely inefficient PMax spend. This shortfall eventually prompted Google to issue credits for a large portion of PMax lead generation campaigns for Brunner clients, and we imagine others as well.
“One crucial lesson from testing PMax lead gen campaigns is the importance of having call tracking and/or CRM integration in place, so you can understand lead quality.”
PMax for Shopping: Better Results for Revenue-Based Campaigns
While PMax for lead generation faced setbacks, we saw better results for PMax with revenue-based accounts running shopping campaigns. Revenue from purchases cannot be cheated on or faked. That means advertisers should carefully consider their campaign revenue goals and choose the appropriate PMax bid strategy.
With those tactics in mind, our team has scaled Shopping Ads programs efficiently by tapping into the additional digital real estate that comes with PMax campaigns. For example, a retail advertising client switched from Standard Shopping Ads to PMax. The client’s revenue more than doubled and their return on ad spend improved. That’s because the PMax optimization algorithms leveraged the rich conversion history of Shopping Ads to find more customers in other marketing channels, such as YouTube.
Unfortunately, we still experienced some challenges in PMax for shopping campaigns. Google’s decisions – such as removing Smart Shopping campaigns and limiting audience targeting and bid strategy capabilities for “standard shopping” – have left businesses with fewer options.
What could make it better? Our team needs the ability to see exactly which networks the campaigns are serving ads on so we can bid down on lower performance areas.
Key PMax Takeaway: Test, Push Back, and Exercise Caution
A major concern for us as PMax power users is the lack of visibility into channel performance and budget allocation. As an agency, maximizing our clients’ budget is always a top priority. Google has PMax data available internally but has not been willing to share channel breakouts. That leads us to question what they’re not saying — and why they’re not saying it. We wonder if account competitors are bidding, unknowingly and unintentionally, on brand and non-brand terms, leading to rising average CPCs in standard search and shopping campaigns.
Is Google the largest benefactor of increased CPCs? Yes, of course they are. A Google vice president testified at a federal antitrust trial that the company increases CPCs in some instances to meet revenue goals. The lack of data from PMax prevents advertisers from scrutinizing performance while increasing revenue for Google, with marketers footing the bill.
Brunner Recommendation: Rigorously Test Google’s PMax before Implementation
Based on our experiences and those of other early adopters and power users, the Brunner team believes it’s essential to push back against ineffective strategies and advocate for advertising approaches that align with business goals and results. We strongly advise against blind adoption of lead generation through PMax.
We urge advertisers to exercise caution and be aware of potential pitfalls. Thorough testing and informed decision-making are paramount in navigating the complex landscape of online advertising with Google’s PMax.
Need Guidance with Effective SEM? We’re here to help
At Brunner, we help clients from many different industries run effective SEM campaigns that deliver measurable impact. In addition to Google Ads certification across our team, we require certifications in Bing Ads to ensure mastery of top SEM platforms. We guide our clients on the best SEM strategy for their business goals, and we can guide your brand, too. Please reach out — we’d love to chat!
Luke Navickas is an associate director of paid search at Brunner. In his role, Luke is a driving force behind the development and execution of high-impact Google Ads and Microsoft Bing Ads campaigns, meticulously tailored to meet specific business objectives. With over 12 years of experience, Luke manages a diverse portfolio of clients spanning various industries, including retail, e-commerce, rental, B2B, higher education, finance/insurance, health/medical, legal, and home improvement.