Performance Max for B2B: When It Works, When It Doesn't, and What to Do Instead

Most B2B companies should not be running Performance Max.

That is not a hedge. If your Google Ads account is optimising toward form fills or free trial signups on-page, and you do not have offline conversion tracking feeding qualified lead data back into the platform, PMax will make your pipeline worse. It will fill your CRM with spam, train the algorithm to find more of it, and cost you weeks of wasted budget before anyone realises what happened.

We have seen this across enough accounts to call it a pattern. PMax was built for ecommerce. It works brilliantly when Google can see a purchase happen, clear signal, clear value, closed loop. 

Lead generation is a different problem. The conversion that matters (a qualified opportunity, a closed deal) happens weeks or months after the click, inside your CRM, invisible to Google unless you explicitly send it back. Without that feedback, PMax is flying blind. And blind PMax is expensive.

But there are conditions where it works. This post covers when to test it, what has to be true first, and what alternatives exist for the job most people are trying to hire PMax to do.

Where PMax sits in the activation order

Performance Max is not the next thing you test after high-intent search stops scaling. It is closer to the last thing.

The order we follow: high-intent search first, then medium-intent, then competitor campaigns (only when you have landing pages that support them, comparison pages, not your homepage). After that, broad match and DSA with strict guardrails. Only after all of those are either maxed on impression share or hitting CAC ceilings do we consider PMax.

The reason is simple. Every campaign type above PMax gives you more control over where your ads appear, what queries trigger them, and which landing pages receive the traffic. PMax gives you almost none of that. It decides where to show your ads across Search, Display, YouTube, Gmail, and Discover, and for B2B, the non-Search placements without consideration are where problems start.

The feedback loop that kills lead quality

Here is what actually happens when a B2B company launches PMax, optimising for on-page conversions:

PMax finds that it can generate cheap form submissions from Display and Gmail placements. Bots and low-quality traffic fill out forms. Google counts those as conversions. The algorithm sees that Display and Gmail are “working” and shifts more budget there. Real search traffic gets starved. Lead quality drops further. The algorithm doubles down on what it thinks is performing. Your sales team starts ignoring marketing leads entirely.

This is not a theoretical risk. We have seen it happen in accounts where PMax was launched with form-fill tracking as the primary conversion event. Nick, who heads up our Performance Department, says that “the biggest cause of PMax failure in B2B has been tracking issues. If your conversion event is susceptible to bots or spam form submissions, PMax will lock onto that traffic and feed the algorithm more of it.”

The industry has started calling this the feedback loop of doom, and it is hard to break once it starts. By the time you notice lead quality has dropped, the algorithm has already learned to optimise for the wrong signal.

What has to be true before you test PMax

If you are going to test PMax for B2B, these are non-negotiable prerequisites. Skip any of them and you will waste budget.

Offline conversion tracking, connected and feeding data. This is the single most important requirement. Your CRM (HubSpot, Salesforce, whatever you use) needs to be integrated with Google Ads so that downstream lifecycle events are sent back to the platform. Not just MQLs. You need SQLs, opportunities, or closed-won deals flowing back with assigned values.

The mechanics also matter. When a sales rep qualifies a lead, that status change should automatically update in Google Ads with a higher conversion value. When the lead progresses to proposal stage, the value increases again. Over time, Google starts to see which clicks lead to real revenue and optimises accordingly. This only works if the data is flowing consistently and the lifecycle stages in your CRM are clearly defined. Messy CRM data will train the algorithm on noise.

Sufficient conversion volume. PMax needs data to learn. The commonly cited minimum is 30 conversions per month, but for B2B, optimising for downstream events like SQLs, the realistic threshold is higher. If your account generates 10 SQLs a month, PMax does not have enough signal to optimise reliably. The learning phase will be erratic, and you will not be able to distinguish real performance from noise.

Brand exclusions. PMax will cannibalise your branded search traffic if you do not explicitly exclude brand terms. Campaign-level negative keywords are now available, so definitely use them. Run brand in a dedicated search campaign where you control the messaging and the landing page. PMax should be finding net-new demand, not eating traffic that was already coming to you.

reCAPTCHA on every form. If PMax is going to send traffic to pages with forms, those forms need spam prevention. reCAPTCHA v3 at minimum. Without it, bot submissions get counted as conversions and the algorithm starts chasing more of them.

A placement exclusion list. PMax serves ads across Display, YouTube, Gmail, and apps. Many of those placements are low-quality, think kid games, bot-heavy apps, made-for-advertising sites. We maintain a standing exclusion list of known junk placements that gets applied to every account as standard. Without it, PMax will happily spend your budget on placements where no B2B buyer has ever made a purchasing decision.

URL expansion turned off (or tightly controlled). By default, PMax can send traffic to any page on your site. For B2B, this means blog posts, careers pages, support docs, none of which are set up to convert a cold click into a qualified lead. Restrict final URLs to your conversion-optimised landing pages.

Where PMax actually works: product-led growth signals

There is one scenario where PMax becomes genuinely powerful for B2B, and it is underused. Product-led growth companies with in-product activation events. 

If your product has a free trial or freemium tier, and users take a meaningful action inside the platform before becoming a paying customer like installing a tag, connecting an integration, completing onboarding, inviting a teammate, that activation event is a far better conversion signal for PMax than a form fill.

Here is why. A form fill tells Google that someone typed their email address. An in-product activation event tells Google that someone signed up, logged in, and did something that correlates with conversion to paid. The gap between those two signals is enormous. PMax optimising for “installed the tracking tag” is working with a signal that is three or four steps closer to revenue than “submitted a demo request form.”

The setup requires passing these in-product events back to Google Ads as offline conversions with appropriate values. It is more technical to implement than standard form tracking, but for PLG companies with sufficient trial volume, it changes PMax from a spam machine into a channel that can identify users with genuine product intent across Google’s network.

This does not eliminate the need for the other prerequisites, you still need exclusion lists, brand exclusions, and reCAPTCHA. But it shifts the quality of the signal PMax is learning from, which is the variable that determines whether the campaign works or fails.

Audience signals are not targeting

One of the most common misunderstandings about PMax is that audience signals do not restrict who sees your ads. They are suggestions to the algorithm for where to look first. If PMax decides that users outside your audience signals convert better, it will serve ads to them instead.

For B2B with narrow target audiences, this matters. You can upload your customer list, add in-market segments for business software, layer on custom segments based on competitor search terms, and PMax can still ignore all of it if the conversion data points elsewhere.

This is not a flaw in your setup. It is how PMax is designed. Audience signals are educational input for the algorithm, not guardrails. The actual guardrails are your conversion tracking and your bid strategy. If your conversion data is clean and pointing at real qualified outcomes, audience signals help PMax find more of the right people faster. If your conversion data is garbage, no amount of audience signal refinement will fix it.

What about AI Max for Search?

Google launched AI Max for Search in 2025, and DSA campaigns are being forcibly migrated to it by September 2026. It does something similar to what many people are trying to get PMax to do, it finds long-tail and conversational search queries you are not explicitly targeting but it stays on the Search network only. No Display, no YouTube, no Gmail.

For B2B, this distinction matters. The long-tail query capture that PMax offers is real, particularly as search behaviour shifts toward more conversational queries driven by AI Overviews and LLM-style searching. But PMax wraps that capability inside a cross-channel campaign that spends across inventory where B2B lead quality is consistently poor.

AI Max gives you the query expansion without the inventory risk. It analyses your existing keywords, landing pages, and ad assets to match your ads to relevant searches you have not bid on, including the longer conversational queries that are growing 1.5x faster than traditional shorter queries.

We are still in early stages of testing AI Max and the results are mixed. We would not call ourselves advocates yet, more like sceptics who keep testing because the product is developing quickly and the use case for B2B is logical even if the execution is not fully there. But if the job you are trying to hire PMax for is “find me queries I am not currently covering,” AI Max is worth testing first because it keeps your budget on Search where intent is highest.

The honest summary

PMax can work for B2B. But the conditions are narrow, the prerequisites are strict, and for most accounts there are better options to test first.

If you have exhausted search, broad, and DSA. If you have clean offline conversion tracking with sufficient volume. If you have brand exclusions, spam prevention, and placement exclusions in place. If you are a PLG company with meaningful in-product activation signals. Then test PMax.=

If any of those conditions are not met, your budget is better spent on the campaign types above PMax in the activation order, or on platforms outside Google entirely. LinkedIn prospecting, Meta top-of-funnel, both are likely to do a better job of demand generation than PMax running on dirty data.