Google Ads Performance Max: maximize ad campaigns by listening to people with machine learning

Have you ever heard of Google Performance Max? This year, Google launched this new tool in beta testing to try and address the post lockdown needs of both businesses and people. We at TSW are partnering with them in this testing activity and we’re going to tell you how we are using it and integrating it into our approach.

Performance Max by Google Ads: why do we use it?

At TSW we have a main objective that cuts across all areas of work: listening to people in order to improve their experiences, whatever they may be. A part of the experiences we deal with take place online and refers to the phase before the purchase, whether it is a simple interest or a decision that is about to become concrete. In this part of the “funnel”, or more simply called purchase path, that becomes more and more focused and narrower (like a real funnel), a key role is played by Digital Advertising on Google and its entire network.

As always, advertising must find a way to capture the attention and remain imprinted in the minds of those who watch it, as well as stimulate the click for purchase. But as we all know, this is increasingly difficult, due to the information overload that characterizes, not only the web, but the whole everyday life of people. For this reason, the solution is to try to talk to the right people according to the message we want to communicate with our Advertising campaigns, namely interested or potentially more inclined to listen. Therefore, we are always looking for tools that allows us to reach, in an increasingly accurate way, these people. This year, an opportunity came with Google’s new tool: the Performance Max campaigns, available from the first half of 2021, for a test phase reserved only for certain clients called “tester clients“.

The birth of Google Performance Max: listening better to people with machine learning

According to the 2020’s “Google Mobility Trends” research, consumer behavior has seen a rapid change, compared to the pre-Covid19¹ period, recording an increase in online shopping for 70% of people. Moreover, it has been analyzed how, for certain products, are recorded more than 500 online touchpoints before making the purchase. This complex path has led to an increase in the acquisition costs (CPA) of digital customers, with a consequent impact on the profit margins of companies in this new digital economy. For this reason, Performance Max campaigns are taking the field, which maximize performance across the entire Google advertising space (Display, Search, YouTube, Discover and Gmail networks, and in 2022 also on Shopping) using machine learning.

Achievable goals with Google Performance Max

As the name suggests, Performance Max’s main objective is to get the maximum efficiency from Google Ads and it does this by offering a wide range of sponsorship targets, which must be aligned with business and marketing objectives. Specifically, the campaigns respond to sales (online and offline) and lead generation objectives:

Google and TSW: a report of the first months of Performance Max

Since May 2021, we have been testing these campaigns for 4 of our clients’ accounts, who have entrusted us their media budgets. In this beta test period, some possible advantages and disadvantages have already emerged.

ADVANTAGES

• Once the goal, or goals, are defined, the machine learning identifies future customers on Google properties, finding the right person at the right time.
• You can help speed up campaign learning through suggestions of audience segments (indicators)
• The algorithm has full visibility of the data since the ads appear on the Google network.
• Optimizations are automatic, which saves time on this once manual operation, and allows you to focus on other strategic factors (such as the updating and the improvement of assets).

DISADVANTAGES

• The control required by the Advertising specialist is minimal: it is possible to set only the assets and to suggest audience segments historically more performing or more in target. It is not possible to monitor “where the advertisements are published”, adjust the offers for the audience segments or clean up the search terms for the “search” positioning.
• At the moment it is strongly recommended to invest, only on this specific campaign, a minimum budget of 10,000€/month so it will be able to record enough data to optimize itself.
• It is necessary, for an initial period, to keep the other campaigns active (search, display and discovery): the question of the budget always comes up, which must be sufficient to cover all the campaigns.

A necessary best practice considering this evidence is linked to the optimization of the active campaigns of the other available types: the ADV specialist must dedicate himself to this activity, which involves a considerable investment of time but is necessary to ensure that the Performance Max campaign learns from the results of the others.

Performance Max results: are new campaigns really designed for people?

Only after testing these campaigns and analyzing pros and cons could we have been able to answer the great question: do they work?

The answer is a definite “Yes“: in fact, in the first months of testing on a finance client with a “lead generation” objective, we recorded, for all accounts where it was activated, a lower Cost per Lead compared to not-brand Search campaigns, which are second in terms of CPL after Search – Brand.

This first successful outcome made us curious to experiment these campaigns with a purchase goal, also exploiting the “shopping” positioning, which is not yet available at the moment.

After testing the effectiveness of Performance Max, what is our assessment? We are very satisfied about effectiveness, the actual results obtained and the excellent improvement in time spent. All that’s left to do is to return this experience to clients and companies who want to get the most out of Google Ads.

¹Source: Google Mobility Trends (2020); Digital Commerce 360, Klaviyo covid-19 poll (Mar 2020); US Department of Commerce (2019)