Google Ads Performance Max: $114 CPL vs Search $126 over $32M lifetime, on 8% of budget

Performance Max has a reputation problem among people who run Google Ads for a living. It hides the search terms. It hides the placements. It reports a blended number across search, display, video, Gmail and Discover, and you cannot see which one did the work. Most practitioners treat it as a necessary evil, something to give a small budget to because Google keeps asking.

Here is what it did in one account over its lifetime, against the channels people trust more.

Performance Max was the cheapest channel in the account. Not by a little. It beat Search (the channel every strategy document is built around), by 10% on cost per lead, over $2.7M of spend and 23,509 conversions.

The recent picture is sharper

The last ten days, by campaign:

The main Performance Max campaign ran at $89 per lead. The best search campaign in the account, on its best-ever bidding configuration, ran at $202. Performance Max got 8% of the budget.

The objection, taken seriously

The standard objection has three parts, and each deserves a straight answer.

"It's cannibalising brand search." Possibly some. Performance Max will bid on brand queries unless you exclude them, and brand queries convert cheaply. In this account, brand was consolidated into its own search campaign with a shared negative list, which limits the leak but doesn't eliminate it on the PMax side. The honest position: some fraction of the $89 is brand traffic that would have converted anyway. What fraction is unknowable, which is the real complaint.

"It's remarketing dressed up as prospecting." Also possibly some. Performance Max uses audience signals and will happily re-serve people who already visited. Remarketing converts cheaply. Same problem: you can't see how much.

"The leads are worse." This one is testable. In this account, the Performance Max campaigns were feeding a lead form asset that had no webhook to the CRM, 13 of 27 submissions in one week came from PMax and none reached the sales team automatically. That's a plumbing failure, not a lead quality finding. Once the plumbing was fixed, the qualified-lead upload rate for PMax leads was not measurably different from search. Small sample, but no evidence of the thing everyone assumes.

Opacity is not the same as performance

Here is the argument, stated plainly.

The complaint about Performance Max is that you can't see inside it. That is true. It is a legitimate operational complaint. It makes the channel hard to diagnose, hard to improve, and hard to defend to a client who asks "where did this lead come from."

But it is not a performance complaint. A channel that produces leads at $89 while hiding its search terms is still producing leads at $89. The lead exists. The cost exists. Whether you can see the query that produced it changes how you feel about it, not what it cost.

Practitioners conflate the two because the first makes them uncomfortable and the second is what they're paid for. If the job is cost per lead, the channel is doing the job.

What would change the conclusion

None of those tests had been run. The channel was being kept small on the strength of an assumption.

What to do

Run the brand exclusion. Google added brand exclusions to Performance Max in 2023; apply your brand list and watch what happens to CPL. If it holds, the objection is answered. If it doubles, you've learned something worth knowing.

Feed the CRM back. Upload qualified leads by click ID for every channel, not just Search. Then compare qualification rates. That is the lead-quality test, and it takes a month.

Scale it in steps. Raise the PMax budget 20% at a time and watch CPL. The channel's cost curve is unknown until you walk it.

And stop treating "I can't see the search terms" as a finding about performance. It's a finding about visibility. Both matter. They are not the same thing.