Push Performance To The Max While Avoiding The Pitfalls


Acknowledge The Flaws, But Embrace The Potential Of Google’s PMax. We’re Not In The “Good Ol Days” Of Straightforward Search Advertising Anymore

The digital marketing landscape is constantly evolving, demanding ever-smarter solutions to engage audiences and drive results. Gone are the “good ol days” of straightforward search advertising being able to support all your online objectives. While search is still a vital foundation to many successful digital advertising strategies, we have officially entered the era of Performance Max campaigns – Google’s game-changing approach that leverages machine learning to personalize and optimize ad delivery across its vast network.

At Smith, we have seen Performance Max’s potential to revolutionize how businesses reach their target audiences and achieve their strategic goals. Just one recent example comes from a set of industrial lubricant ecommerce campaigns focused on driving sales and new-customer acquisition. In the first quarter, with Smith managing the account, we’ve seen a YoY ROAS increase of 23% resulting in $2.3M in incremental conversion value vs. the same period last year. 

The key was this campaign used our ONER method to take back some of the control you relinquish to Google by running PMax in the first place. It leveraged the learnings from the previous months, but targeted a more highly segmented and relevant audience. Within the first month we identified asset combinations that produced more optimal results. We’re not the only ones seeing this type of success, Google has published numerous case studies on the topic of PMax as well.

On the flip side, this campaign type has caught the ire of many knowledgeable and trusted digital marketers concerned with Google’s “black box” approach to certain data components within PMax. There’s also concern over the limited controls offered to advertisers in terms of placement allocation. Not to mention less predictable spend/performance patterns. 

They’re not wrong. 

PMax, just like automated bidding, smart-shopping and countless other technological advancements before them that most search marketers have now come to trust, is full of potential pitfalls.

However with the promise of performance so great and the risk of unknowns seemingly overwhelming, how are you supposed to proceed? You know you can’t just bury your head in the sand and wait for it to be perfect while competitors pass you by; and while it worked out for Indiana Jones In The Last Crusade we don’t recommend you just take a giant leap of faith either.

Instead, you need to develop a strategic, data-centric approach, with an incremental testing and optimization roadmap specific to PMax campaigns. We call it ONER a British word for 1 of 1 and it stands for: Optimize. Navigate. Evaluate. Repeat. 

Here’s the basis of the “not so secret sauce” approach we take any time one of our brand partners engages us on what to do about PMax.

Optimize: Always Ask Who You’re Trying To Reach & Why You’re Trying To Reach Them

Audience segmentation is at the heart of every successful PMax campaign. Many marketers, focused on the lack of insight into specific placement successes, fail to realize that the initial premise of the campaign is to focus more on people than placements. The goal has always been to serve ads to the right people wherever they’re engaging with content, as opposed to a single more highly controllable channel. As long as it gets you the desired conversion action, does it really matter the exact channel credited with the conversion?

So ask the questions: who are you trying to reach? What are the terms they’re searching based on the actions you want them to take with your brand? Leverage your own search KW data, the Google KW planning tool and 3rd party tools like SEMRush to develop this list. Then mix in competitive analysis of SERPs, Merchant Center rankings for competing advertisers of similar sizes and market insights from your brand partners. Finally, research what apps these competitors have as well as complimentary apps based on the desired actions you want consumers to take. In 2024 Google will “upgrade” to Themes instead, but you can still duplicate older asset groups to take advantage of these custom search intent options.

Navigate: Use Data-Driven Decisions To Define Success Metrics

Speaking of those desired conversion actions, establishing the right goal inputs is pivotal to campaign success. With current search, shopping and YouTube campaigns + GA4 or other analytics tracking platforms, you’re equipped with a wealth of data that can inform campaign KPIs. Even with challenges in getting some of the data, once you’re running Performance Max, you’ll have access to campaign specific insights, revealing the effectiveness of your individual assets and the overall campaign. Clicks, conversions, and engagement metrics become your guiding light, illuminating your strengths and highlighting potential areas for improvement.

Use a mix of transactional and revenue driving activities with offsite conversions as needed. For example with retail customers we often deploy a local-actions version of the campaign prioritizing store visits and directions requests + a commerce-oriented campaign prioritizing add to carts, app downloads, calls and purchases. We give our commerce campaigns the additional goals to help define target audiences better, but we bid to target ROAS and only utilize purchase revenue for the value so each asset group will try to get any of the listed conversions it can, but spend is directed most toward users likely to purchase.

Evaluate & Repeat: Personalizing Creatives Based On Search + Continually Evolving Through AI-Powered Tests

Performance Max harnesses the power of AI to elevate your campaign to new heights. We feed the platform keywords and descriptions, and let it automatically generate compelling headlines and even images for specific testing asset groups vs. the controls that consist of our top performing search headlines and descriptions. 

Using a mix of AI-generated assets in this limited capacity vs. fully handing over the keys for all asset groups, keeps us more in control of the key messaging components we want to deliver to consumers, without missing out on the vast library of creative possibilities offered by AI. It also accelerates our ability to test, learn and develop better performing asset groups. PMax campaigns are constantly analyzing asset and ad group performance metrics to deliver the best results. In set periods of time, we harness this performance data, put the best performing assets into the control asset group and “reset” the test with new messaging, AI opportunities and even target audience segments.

This approach helps us limit performance spikes caused by AI “learning” and ensures that even if the AI outperforms the original asset group, we’re still improving our original asset groups and putting the machine learning right back to work on improving again. This iterative process maintains our ability to direct budget to the best performing variations just like we would a specific variation of a search KW in the “good ol days.” So maybe they’re not that far gone after all.

Ready to unlock the full potential of Performance Max for your brand?

If you’re interested in learning more about how Smith is generating efficiency gains and performance-driven results from Performance Max campaigns and other AI-powered campaigns & tools, Let’s Talk.