17 July 2024

Google silently release an ad targeting revolution

In summary

• Customer Types are a new feature within Google Ads that allow you to inform Google’s algorithms of the audience category your customers fall into
• There are currently nine Customer Types and two have already been integrated with existing bid strategy goals
• The release of this feature, in conjunction with others, could suggest that Google is moving away from keyword targeting into a ‘keywordless’ future

Customer Types quiet release onto Google Ads

In what seems like the shadow of the night, Google stealthily implemented their newest feature, Customer Types. This is the latest in a suite of releases; accelerating advertising on Google Search away from its traditional keyword bidding origins towards machine learning and “AI” based bidding and audience focused targeting.

What are Customer Types and what do they do?

Customer Types allow you to label your audiences as one of nine different types of customer. While upon first glance that doesn’t appear too exciting, if we consider changes made to Search Ads in recent years, the value of this feature begins to unfold.

The nine available Customer Types at the moment are:

• Purchasers
• High-value customers
• Disengaged customers
• Qualified leads
• Converted leads
• Paid subscribers
• Loyalty sign-ups
• Basket abandoners

How do they fit in with other Google Ads AI features?

Customer Match lists are already used by Google as similar audience signals in the background, as a black box replacement for the previous similar audiences feature. As with other black box features such as Performance Max (PMax), the exact mechanics behind them are a secret within Google. Despite that, we can speculate on their workings. For example, we know PMax campaigns require some level of human guiding and intervention, as seen with the Search Trends feature. Doing so, provides PMax with audience signals that help expand its reach.

Deriving our understanding from Search Trends, alongside Google’s documentation, we can theorise that Customer Types provide bidding strategies with a baseline understanding of your audiences, as opposed to them learning over time. In essence it’s like giving your bid strategies a head start on performance.

This isn’t all however. Customer Types will also integrate with soon to be released hyper-specialised bidding strategies. As for right now, new customer acquisition integrates natively by excluding all customers in the “Purchaser” audience, and “High Value Customers” being used as a signal by maximise conversion value strategies.

“As Google releases additional customer lifecycle goals beyond new customer acquisition, such as re-engaging disengaged customers, labelling your audience customer types will become even more important.”

From this, we can assume an expansion of bidding strategies targeting/interacting with all nine of these Customer Types in the future.

The many changes to search ads over the years, and what they suggest

Match type rules relaxation

The definition of an “exact match” keyword has become murkier over time, particularly with the release of “close variants”. This allowed plurals, misspellings and other very similar search terms to your exact keywords to be able to trigger an ad. The list of what was considered a “close variant” was ever-expanding in the following years, to the point where by 2018 phrase match was considered more exact, than actual exact match.

With the deprecation of Broad Match Modifier (BMM) at the beginning of 2021, the same loosening of exact, came for phrase, as much of BMM’s targeting was incorporated into phrase match.

What this meant was that advertisers had to become even more vigilant of their negative keywords, no longer able to rely on strict keyword setup to prevent them from showing on irrelevant queries.

The Broad Match push

In early 2023, Google began their continuing push towards the adoption of new Broad Match. With it Google tried equalising the previously maligned Broad Match with its phrase and exact siblings. This was done by providing it with additional bidding signals:

• Landing pages
• Other keywords in ad group
• Previous searches
• User location

The change seems to have paid off, as in our experience Broad Match tends to outperform phrase, and equal or perform slightly worse than exact keywords across an account as a whole.

The table below shows the signals that are available for the different match type options.

Signals considered to drive performance	Exact and Phrase	Broad Match
Keyword	Yes	Yes
Landing pages	No	Yes
Other keywords in ad group	No	Yes
Previous searches	No	Yes
Predicted performance	Yes	Yes
User location	No	Yes

AI is coming for search too

Alongside these changes to Broad Match, we have also seen the release of minimal control, machine/AI based campaigns such as PMax and DemandGen that work across multiple ad formats (not just search). In the case of PMax, this campaign type has had such strong performance for some, that they have pivoted almost all of their Google Ads spend into this campaign type. If PMax begins to consistently outperform traditional search campaigns, performance advertisers will shift their spend, which could leading us into a situation wherein all campaigns become AI based.

The launch of Power Pair

During Google Marketing Live 2024, it was announced that Google’s new ideal Google Ads setup had become even more automated with the reveal of Power Pair. This new feature is a combination of PMax and Broad Match keywords. Effectively the power of PMax is further strengthened through insights and signals provided to it by these Broad Match keywords.

In this Power Pair future, Broad Match’s role continues to be that of finding new customers. However, instead of working as a phrase and exact match keyword generating resource, it now finds new audiences that are then leveraged by PMax campaigns.

If the performance provided by Power Pair is as strong as Google suggests, many accounts could begin to transfer to this method over the next year. A setup implementing lower funnel/conversion generating search campaigns via Pmax, with Broad Match search campaigns fed into the PMax campaign signals.

Brand exclusion and brand restriction features

Almost from the offset, PMax’s strong performance figures were suspected to be potentially derived from cannibalising brand campaign terms by advertisers. Google responded to this criticisms with brand exclusions and brand restriction features. This differed from excluding brand keywords from PMax by excluding brands based upon their urls.

So, instead of excluding “Amazon” as a phrase keyword, you would exclude the website “https://www.amazon.com.au/”, and all the terms associated with that brand considered by Google, further shifting even negative keywords away from their manual upbringing into its automated future.

Google has also recently released brand inclusions for broad keywords and performance. Although it is currently too rudimentary to rely on based on our tests, with improvements we could potentially see both negative and regular keywords being deprecated entirely.

Third party cookie deprecation

Third party cookie deprecation continues to be the big news in digital advertising, the Boogeyman still lurking in the hills of 2025 or beyond. Regardless, third party cookie deprecation and the shift by Google towards these first party data based solutions are interlinked with this new Customer Type feature.

The now deprecated Similar Audiences, required third party cookies, and were replaced by Customer Match audiences. Using this feature, a user would upload the following pieces of personal information about a customer:

• First name
• Last name
• Email
• Postcode
• Address
• Phone number
• Country

This then links into the Google’s Signals feature whereby this information would connect actions with users logged-in behaviour via their Google account. This connection between consumer behaviour on your website and the Google account is one of the most prominent of a handful of features allowing for remarketing bid adjustments to work when third party cookies do eventually fade away.

The other solution for remarketing is the Topics API. The Topics API essentially tracks all Google user’s behaviour across websites and allocates them into certain audience groups, but only one at a time. These audiences will over time become more and more stratified and specific. It differs from the existing affinity and in-market audiences as instead of directly tracking website visits, the Topics API only knows the types of websites that the person visits, i.e. car websites versus clothing websites.

So, what is the future of Search?

Using all we have learnt, Customer Types can be seen as one of many leading the charge from keyword to audiences based targeting, propelling a search campaign’s setup closer to display and video implementation. Instead of dividing campaigns by keywords, they are instead divided by the audiences to who we want to target.

With the the power of exact and phrase keywords being eroded and their performance being hindered more and more, they could easily fall behind that of Google’s Broad Match golden child. And the question remains - how long before they, as BMM before them, are deprecated due to lack of use?

All this suggests a potential future where your setup could be as follows:

• PMax campaigns for lower funnel and conversions
• Broad Match campaigns targeting new customer acquisition for data collection
• DemandGen for upper funnel activity
• Additional PMax campaigns on the side using the new bid strategies that Google has hinted at - disengaged customers, qualified leads, and basket abandoners

The diagram below shows what the potential future of Search/Google Ads digital advertising funnel might look like.

With the upcoming third party cookie deprecation, Customer Match is being rapidly adopted. Through these first party data solutions, the need for Broad Match may diminish entirely as Google’s Topics API advances. It would allow Google to use your existing first party customer match data to identify topics based audiences that have strong performance for your campaigns, thereby cutting out the middle man of Broad Match. This in turn, finally leads us to a world where only AI campaigns are available.

Louder’s recommendation

The opinions shared in this article have been formed based on the information and trends that we’ve observed from Google around search campaigns in recent years. Our speculation of what’s to come has not been confirmed by Google.

Louder’s recommendation for Customer Types is:

1. Keep an eye out for an official feature release by Google before trying to test it. Whilst it is available within the interface, it is still in a buggy state at present.
2. Test the new features. Despite some continuing reservations, PMax has generated huge volume and performance gains for many. Alongside your testing, make sure that you maintain elements of your existing structure to fall back on in case these new features incur any issues.
3. Begin building more durable Customer Data Platform (CDP) consolidation and integration solutions to support all the new privacy centric feature releases. Companies are faced with the need to ensure they have accurate, valuable, and easy to access first party data, or risk being left behind as Search evolves from keyword to audience based targeting.

Need help with your Search setup?

As Search strays further and further from its roots, it can be hard to keep on top of what features to use, and which ones to discontinue.

Get in touch with Louder to discuss how our specialist search team can help you or sign up to our newsletter to receive the latest industry updates straight to your inbox every month.