Our team attended a session hosted by Lunio about AI Max. If you’re unfamiliar, AI Max is a new Google Ads feature that allows you to opt campaigns into AI-settings. It acts as a Dynamic Search Ads (DSA) style campaign, with the ability to opt into ad copy customizations, extends search query matching, and can leverage landing page expansion. Google pitched AI Max as Excalibur; we got a butter knife with glitter.
Jokes aside, the rollout hasn’t been all bad, new product launches are always mixed bags, and we still remember the PMax beta. While we’ve come a long way from spam leads and poor quality traffic, we are finding AI Max isn’t adding as much value as originally touted by Google, and still needs quality refinements, with Lunio citing a 35% increase in invalid traffic volume upon launch. In a split test where we turned on AI Max for 50% of nonbrand traffic, we found that the AI Max enabled campaign fell behind the control in terms of lead volume by 5% and cost per lead by 2%. AI Max contributed a measly 12% of leads on 20% of clicks from the test campaign, which might not be bad if its purely incremental volume, but the types of queries AI Max was matching to raise suspicions on quality. We only tested one arm of AI Max: search term matching. And while Google has stated full adoption (search term matching + copy creation + landing page expansion) will generate the best results most of our clients aren’t willing to give up that much control to the Google machine.
Our original hypothesis was…are we already covering our bases so well that Google just can’t find any-value add? While the initial thought may sound a little self-serving, we were not the only ones with similar remarks. Remember that Lunio panel? The TL;DR for a lot of the panelists was PMax is going broader than whatever your current keyword matchtype is set to, and if you’re already on broad, it may pick up additional volume elsewhere, but that traffic is often not converting, and then Google learns and pairs it back.
So, where does that leave us?
At LQ Digital we are grounded in a test-and-learn strategy, and we are always challenging ourselves to continue to push performance through new tests, so we will keep running and trying, and performance isn’t hurting, but it’s probably also slipped in testing priority across accounts that haven’t already opted in. If you’re testing in AI Max and finding you’re gaining a lot of new ground, it may be time to revisit your search strategy and account coverage and you’re one click away from the contact us button if you need a fresh perspective.
