Thought on Google AI Max?

[wihanvanderwalt]

In my testing, it hasn’t really worked for my clients. I noticed a lot of traffic came from “ideation” or “ideas”-type queries — people looking for inspiration rather than taking action. That kind of top-of-funnel intent just doesn’t convert well, so performance dropped off compared to more controlled setups.

[jwiegand]

I’ve found it works great when you have multiple markets, because some search terms that are common in one market are different in others and it helps to cover that gap.

I’ve been using the Experiments to run it on different countries with mixed results, in one market it dropped performance (conversions) by 25% and in other it increased by 125+%

[potatodrinker]

Australian tech company, in-house PPC team of 4.

We run AI max through an A/B experiment (campaign vs with AI max on) across half a dozen campaigns. Conversions are about +20% on control, CPA is between 10% better to 10% worse so overall better than I expected from what is essentially Broad Match on steroids.

We’re 4 weeks into the 6 week test but don’t expect results to differ much. 80% confidence interview puts the conversion uplift as a real deal.

Onto problem is that the early days had a lot of irrelevant keywords come up that needed heavy negative work

[Suitable-Matter9339]

Tested in some accounts of our clients. Waste of money. Looks like a “extra broad” feature.

[SeboFiveThousand]

Most accounts it was objectively worse than broad match on smart bidding, have seen a handful of successes from higher spend + high conversion generic search campaigns so I’m assuming you need to be at a mature state for it to make sense to open up the targeting

[benl5442]

Its worked well for me when it’s clear the intent from keywords.

However when there is a line between business and consumer, it struggles, like energy switching. It just goes crazy for consumers when you want business.

[Sea_Appointment8408]

All it’s done is waste spend for me and make things worse

[PleaPleaCee]

I use it for hotels but have expanded keyword discovery off. It often started serving restaurant or wedding terms. Location of interest is what I was most interested in anyways

[AdOptics]

It takes a TON of data to reach any kind of statistical significance in a/b testing. The outcome is a conversion. Unless you have thousands of conversions, there is no reasonable way for Google to effectively test all broad matches (along with every ad asset combination) and have it be an actual conclusive test.

[Crazy-Car948]

Absolute trash

[Revolutionary_Sir393]

It’s broad match on steroids, you’re handing the keys to google and letting them spend all your gas hoping they pick up leads along the way

[Few_Presentation_820]

For lead gen, I haven’t seen any impressive results yet compared to a standard search campaign. Small accounts who can’t afford to test should keep a distance as AI max isn’t there yet

However, it might slightly better the CPL in larger account using a lot of broad match keywords & a ton of past data.

It’s also likely to get better overtime as google trains on more advertiser data points just like P max did

But as of now, it’s either a 50/50 experiment or a big no if the ad spend is small. The reason is, AI max is nothing new but broad match + DSA packed under one shinny button so that means losing control which google loves

[PPCNotPCP]

I didn’t think it could get worse than broad match. Google proved me wrong again.

(Also work in healthcare advertising)

[Viper2014]

Some clients were really interested in trying them out. The results were absolute trash and all of them asked for the campaigns to be turned off.

No suprises there : )

[liljayman93]

AI max gave us so much irrelevant traffic that I turned it off after 1 day. A Google higher rep reached out & said that it takes time to optimize. We done nearly over $1m on google ads with so much data and negative keywords now, don’t know why we would need to further optimize for a new feature to work lol

I run a local business & Wish Google would just give us the leads that we want and advertising for instead of coming up with new “features” to generate expensive nonsense traffic.

Sorry for the rant 😅