Is Your Marketing Chasing Its Own Tail?

If you feel like, “Google just doesn’t work for me, or at least not how it used to,” you aren’t alone.

What you may not realize is that your media buyer isn’t who you think.

Your Google media buyer is now AI. This is a fundamental shift that requires new strategy, management and measurement to drive results. Otherwise, your marketing campaigns will end up chasing their own tails.

Here’s the good news. Google still works, and it can work for you.

Here are the keys to success:

1. Understand how AI campaigns work.
2. Build off of strategies proven to be incremental at scale.
3. Use independent measurement to drive better management.

"PMAX and AI MAX are not campaigns with AI features. They are Google AI-driven products that happen to run ads. The advertiser sets a budget and a goal. Everything else is the machine."

Understanding how AI campaign types work is essential to manage them correctly. PMAX, and AI MAX are fully automated systems. Google's AI controls bidding, audience expansion, creative assembly, placement, and attribution simultaneously.

Recent transparency improvements (asset-group reporting, search term insights, brand exclusions) give advertisers more visibility, but the core execution is automated.

AI runs the campaign and generates the performance data. There is no separation between execution and measurement.

Your new media buyer, AI, that optimizes for reported metrics is also the one reporting them. 

As a result, brands are leaning out of non-brand AI campaign types

Either by fault of set up or visibility into performance, across the 138-brand dataset, the response to AI campaign types has been consistent.

Brands are shifting budgets from non-brand AI campaigns to more comfortable, manual formats that keep humans in the loop.

For instance, non-brand shopping gives keyword-level control while non-brand PMax gives the AI full autonomy. The money is moving accordingly. Intentionally or not, brands are moving money to where it feels comfortable and they can stay in the driver’s seat.

• Google Non-Brand PMax, the largest AI-controlled format, is contracting.
  • -8.5% YoY
  • -10.8% Q1 vs Q1.
• Google Non-Brand Shopping, the format with keyword-level human control, is accelerating.
  • +3.2% YoY 
  • +17.7% Q1 vs Q1.
• Google Brand PMax, where advertisers are more comfortable letting AI run when it is working a controlled, branded signal.
  • +22.1% YoY 
  • +28.2% Q1 vs Q1

What hasn’t changed

"Platform reporting is generated by the same system running your campaigns. Across 138 brands and 172 geo holdout studies, it overstates real return by 1.75x to 2.97x, consistently, regardless of campaign type."

• Both PMax and shopping: platform ROAS 4.86x vs. tested iROAS 2.05x a 2.53x overclaim
  • 94% of brands in this group overstate
• PMax only: platform ROAS 3.45x vs. tested iROAS 1.88x a 1.55x overclaim
  • 95% of brands overstate
• Shopping only: platform ROAS 5.43x vs. tested iROAS 1.60x a 2.97x overclaim
  • 100% of brands overstate
  • This is likely because many brands use non-brand PMAX and brand shopping leading to higher platform ROAS and lower iROAS.

This is not a finding about one campaign type, one brand, one size, or one time period. It holds consistently true across 15 months of data and every set-up tested.

Question for the room: If you moved budget into AI campaigns, did you also build a plan to verify the AI is delivering what it reports?

Breaking optimization -> measurement circularity is essential to driving performance

Algorithms are very literal. When you tell a campaign to optimize for conversions it can measure, it is going to do just that. Ultimately, marketers can’t blame campaigns for doing exactly what they were told to do, “get attributed conversions.” 

That is the difference between attribution and contribution. Attribution is getting credit, contribution is money in your pocket. 

Independent measurement gives marketers the clarity and confidence to spend on tactics that drive contribution without worrying about trying to get attribution. 

That is the only way to break the circularity of this loop. Put another way, this is how you get your campaigns to stop chasing their own tail. 

The winning PMAX structure, proven by independent measurement

PMAX and standard shopping are better together. Together they tend to outperform PMAX alone. When properly structured, managed, and independently measured, they deliver the strongest, verified incremental revenue together. 

• Both PMAX and shopping tested median: 3.08x iROAS across 84 geo holdout studies.
  • For the bottom 25% cohort: 1.24x iROAS. 
  • For the top 25% cohort: 6.50x iROAS. 

What makes the difference is what you do next:

1. This month: Audit your AI campaign set up.
   • Normalize audits. As a former media buyer, I can tell you we become attached to our accounts and setups as our own special little works of art. What has worked before, and importantly, what has not worked before, changes.
   • Run a setup audit across all active AI campaigns. Audience exclusions, product feed status, bidding strategy, and campaign structure.
   • Start with setups proven to drive incremental results. If you don’t know what they are, ask the team at Measured.
   • The audit does not require new tools or a new budget. It requires looking at the configuration decisions that are already made and whether they match the winning profile.
2. This Quarter: Run an independent holdout incrementality experiment .
   • One test answers three questions. Is your strategy chasing attribution or contribution? Which direction do you diverge from benchmarks? Are you leaving upside on the table?
   • If you get a lower than expected result, don’t be discouraged. That indicated opportunity. Work with your partners at Google and Measured together for visibility into strategies other brands are using to average an iROAS over 3x on google.
3. Ongoing: Make independent measurement standard.
   • AI controls execution, placement, and attribution simultaneously. Independent measurement is the only mechanism that sits outside all three simultaneously.
   • Unify finance and marketing teams to optimize for real business outcomes. Rather than by the numbers generated by the system doing the spending.
     

Google campaigns work. But must be optimized to chase squirrels, rather than their own tails.