For three years, the loudest complaint about Performance Max was the same: you couldn't tell the algorithm which channels mattered. Google decided where your budget went across Search, YouTube, Display, Discover, Gmail, and Maps. Your only levers were indirect: asset quality, audience signals, search themes, and conversion value rules.

That's changing. Google is alpha-testing a "Channels" prioritization control inside PMax. But the mechanism isn't what most advertisers expected, and for B2B SaaS teams running pipeline-based motions, the risk of misuse is real.

What the Controls Actually Do

The new setting lets advertisers apply positive or negative adjustments per channel. A positive adjustment relaxes the CPA the algorithm will accept for that channel; a negative one tightens it. So if you set a positive adjustment on Search, PMax becomes willing to pay more per conversion there, signaling that you value Search-sourced conversions higher.

This is CPA-tolerance signaling, not budget allocation. You're not saying "put 40% of spend on Search." You're saying "I'll accept a higher cost for Search conversions because they're worth more to me." PMax still handles actual spend distribution. That distinction changes what you need to get right before you touch these controls.

The Attribution Trap Waiting for You

Here's where it gets dangerous. A buyer who first sees your YouTube ad and later converts through Search will attribute the conversion to Search. YouTube looks expensive. Search looks efficient.

If you react by tightening YouTube's CPA tolerance, you may be cutting the channel that created the demand Search captured. You'd be optimizing the scoreboard while degrading the system that feeds it.

Platform dashboards show credit assignment, not causal contribution. The channel controls amplify whatever biases exist in your attribution. If your attribution is wrong, you'll steer PMax toward the wrong channels with more force than before.

For B2B SaaS: Conversion Quality Is the Prerequisite

The channel controls are only as good as the conversion signal you're feeding PMax. If you're optimizing to form fills, PMax will find the cheapest form fills across channels. A positive adjustment on Display in that scenario just means more cheap, unqualified submissions at a higher CPA. Worse than the default.

B2B-focused guidance from multiple practitioners converges on the same point: PMax works for pipeline generation only when offline conversions or CRM-imported milestones are the optimization target. That means importing demo completions, SQL handoffs, proposals sent, or closed-won events from Salesforce or HubSpot back into Google Ads. Without that feedback loop, channel prioritization is a lever connected to nothing useful.

The hypothesis is straightforward: if you feed PMax qualified-stage conversion data and apply a positive channel adjustment to Search (where intent is highest), pipeline quality should improve while cost per qualified lead stays within guardrails. But that only works if the conversion events you're importing actually correlate with revenue. If your SQL definition is loose, you'll get more loose SQLs from Search at a higher cost. Garbage in, confidently steered garbage out.

How to Test This Without Wrecking Your Pipeline

Setup: Import at least one mid-funnel CRM event (demo completed or SQL accepted) as your primary conversion action in Google Ads. You need a minimum of 30 conversions per month for PMax to optimize effectively.

Baseline: Run PMax with no channel adjustments for 2–3 weeks. Document channel-level spend distribution, CPA by channel, and downstream pipeline metrics from your CRM.

Experiment: Apply a modest positive adjustment to Search and a modest negative adjustment to Display. The hypothesis: if we increase CPA tolerance on Search and tighten it on Display, then cost per SQL will decrease because Search captures higher-intent demand. Guardrail: total qualified pipeline volume doesn't drop more than 15%. Stop-loss: if CPA increases 25% with no corresponding improvement in SQL rate after 3 weeks, revert.

Readout: Evaluate at the SQL and opportunity level, not at the lead or form-fill level. Compare channel distribution shifts against pipeline outcomes. If Search share increased but SQL rate didn't improve, the adjustment didn't work. That's useful data.

The Uncomfortable Reality

These controls are in alpha. They may ship differently, or not at all. Google's track record with PMax features includes plenty of tests that never reached general availability.

What isn't premature: fixing your conversion data. Whether or not channel prioritization ships, every PMax improvement depends on the quality of the signal you send the algorithm. Teams importing CRM milestones and running holdout-based measurement will benefit from channel controls if they arrive. Teams optimizing to form fills won't benefit from anything Google ships, because the algorithm is doing exactly what they told it to: finding the cheapest conversions, wherever they live.

The control Google is testing gives you a louder voice. It doesn't give you better judgment. That part is still on you.