Initial Results of Google Ads’ Performance Max Campaigns

Google is aggressively pushing Performance Max campaigns in 2022, but advertisers still have hesitations regarding performance, the lack of control and insight, and how this new campaign type will impact strategy moving forward.

Knucklepuck wanted to find out the answers to some of these questions, so we began testing fairly early on. A summary of the key points and findings are below.

What We Tested

In May 2022, Knucklepuck launched Performance Max campaigns for a client in the real estate industry. The primary goal for this account is lead generation through leasing applications and contact form submissions, with a secondary focus on generating brand and property awareness.

We used the following criteria to define the Audience Signal in our Asset Groups:

Custom Segments – audiences defined with popular keywords, URLs, etc. of our target market.
- Ex: Competitor terms
Our Data – audiences defined based on previous interactions with the client.
- Ex: Website visitors and remarketing lists
Interests & Detailed Demographics – relevant audiences from Google’s pre-defined segments.
- Ex: “Apartments for Rent” in-market audience, “Business Professionals” affinity audience

Demographics, such as gender, age, income, and parental status, can also be used to define the Audience Signal. However, we did not use this as part of our criteria because it was relevant to target all demos. This is also Google’s default and recommended setting.

Additional set-up notes include:

Campaign Structure – Campaigns were segmented by property, as the majority of them have different locations and target markets.
Bidding & Budget – Tested using the Maximize Conversions bidding strategy and a daily budget of $10.
Ads – Assets include 20 images, 1 logo, 1 video, 5 headlines, 1 long headline, and 5 descriptions (one of which has to be 60 characters).
Other – We did NOT use dynamic ads, local product feeds, final URL expansion, or ad schedules.

Results

From 5/24/22 to 8/21/22, Performance Max campaigns drove the following results compared to the rest of the account (which includes traditional Search, Display, and Video campaigns):

Note: Micro-conversions = calls + directions, which are important to the client but not a main KPI when analyzing performance.

As shown above, we saw significant improvements across all priority conversion metrics: cost per form submission, cost per action, and conversion rate. In terms of awareness and general efficiency, we saw a higher clickthrough rate and lower cost per click, but the cost per 1000 impressions was much higher as well.

It’s important to note that the low CTR and CPM for the account (non-PMax campaigns) may be somewhat skewed. That data includes Display campaigns, which drove 9.2M impressions compared to 51.6K for PMax.
Additionally, ~60% of PMax conversions over the past 28 days are attributed to branded searches.

Overall, the Performance Max campaign tests were a success, with the caveat that branded searches heavily contributed to these results.

There are also some FAQs we’ve come across in the PPC community since Google Marketing Live in May 2022:

Were impressions, clicks, etc. cannibalized from brand and/or non brand campaigns?
- It is somewhat difficult to answer this at the account level due to fluctuations in monthly budgets and campaign coverage. That said, we were able to reference one property that had the same budget and Search campaigns running from 5/24 – 7/31.
- For that set of campaigns, Search ads only saw a 6% decrease in impressions compared to the previous period. There was also a 22% increase in clicks and 29% increase in impression share.
  - This indicates that the decrease in impressions was likely due to account maintenance, such as adding negative keywords, and not the launch of PMax. However, please note that this data is still limited, so these results may not necessarily be the same across all accounts.
Did overall / incremental performance improve or decline?
- Compared to the previous period (2/23-5/23), the account saw a 50% increase in conversions and 60% increase in micro-conversions. This resulted in a 17% decrease in the cost per form submission and a 20% decrease in CPA. Some of these shifts are due to seasonality and general account maintenance, but it is safe to say that PMax contributed as well.

Limitations & Next Iterations

The first limitation of this test is that the lead quality isn’t very clear. It’s great to see a higher quantity of leads and lower CPA. Results also seem positive based on anecdotal feedback from the client, such as “the sales team is happy with what they’re getting.” That said, there isn’t hard data (like lead scores) to reflect that as of now.

Moving forward, our plan is to get more specific figures related to lead quality and/or conversion value so that we can make sure spammy leads aren’t inflating our numbers.

Another limitation of this test is inconsistent budget levels, as our monthly budget changes each month based on the client’s internal data and priorities. While the budgets for the PMax campaigns stayed the same throughout this testing period, the budgets for Search, Display, and Video campaigns fluctuated significantly. Because of this, it’s difficult to fully determine how much PMax campaigns may be cannibalizing traffic from other campaign types.

In terms of what we want to test next, refining our coverage will be the main focus. These campaigns are currently running without any account-level placement exclusions applied. Other advertisers have also noted that you may be able to negate keywords at the campaign level, but as of now that is only possible if you have a Google representative and ask them to do it for you. Additionally, Google notes that account-level negative keywords “will soon be available.”

In the upcoming weeks, we plan to add account-level placement exclusions. We will also follow up with our Google rep about adding campaign-level negatives.

Conclusion

Historically, we have been cautious when testing new campaign types in Google Ads. This is especially true for ones that are heavily automated and give the majority of control over to Google.

Performance Max campaigns are no exception, but they are absolutely worth testing. These campaigns were a success for our client, so we will continue to run them in this account and begin testing for others.

Lastly, it is important to set expectations with your bosses/clients that this is a new campaign type. Google’s algorithm needs time to learn and optimize. After 2-3 months have passed, you’ll be in a better position to determine whether or not PMax becomes a mainstay in your Google Ads strategy.

As always, feel free to reach out on Twitter with any questions, comments, or feedback!