A $25 cost per lead sounds like a gift. A $2,140 cost per SQL sounds like a problem. Both numbers came from the same campaign type. A $25 cost per lead (CPL) sounds like a gift, while a $2,140 cost per sales-qualified lead (SQL) seems problematic. Both figures come from the same campaign type. In a B2B SaaS benchmark dataset, Google Performance Max (PMax) reported a $25 CPL compared to $143 for standard Search campaigns, with a conversion rate of 6.53%—the highest among measured campaign types. On paper, PMax excels. However, the benchmark has a caveat: PMax conversions may not be directly comparable to Search because it combines branded queries, lower-intent placements, and remarketing inventory into one category. The efficiency is evident, but the quality remains questionable. ### The $6.6 Million Blind Spot A 2026 audit of 104 enterprise B2B SaaS accounts quantified this uncertainty, revealing $6.6 million in wasted spend, with 25% directly linked to PMax campaigns lacking offline conversion tracking. Accounts that fed CRM-qualified outcomes back into PMax saw their cost per SQL drop to $620, while those "running blind" paid $2,140 per SQL for the same campaign type. That’s a 3.5x gap—not due to creative or targeting changes, but because the algorithm lacked clarity on what constituted a good lead. The trade-off when skipping offline conversion feedback is significant: PMax optimizes based on the signals you provide. If you give it form fills, it targets users who fill out forms. If you provide SQLs with revenue data, it identifies a different audience. The system works, but only towards the objectives you set. ### Overfrequency Isn't a Frequency Problem Google is now beta-testing an option for PMax advertisers to exclude Google Search Partners (GSP) and Google Display Network (GDN) inventory. Media buyers are optimistic, as GSP has historically been a source of unsuitable placements that advertisers couldn’t exclude or even see. Until recently, Google required PMax advertisers to run there and withheld placement data. However, this inventory-exclusion fix addresses only one issue. The deeper problem for B2B SaaS teams is overfrequency masked as performance. Cadent President Doug Rozen bluntly stated on the Next In Media podcast that viewers don’t think, "That brand has great frequency," but rather, "Make it stop." While his comment referred to CTV, it applies to PMax as well. When the algorithm focuses spend on remarketing and branded queries, it can achieve your numbers while saturating the same users. Attribution dashboards report conversions, but users experience persistence, not precision. Overfrequency in PMax isn’t about a single frequency cap; it’s about overexposure to the wrong audiences, inflating attributed conversions without increasing new customers. If PMax consumes a large portion of your budget, incremental spending tends to shift toward remarketing rather than new acquisition. You may appear efficient while actually shrinking your addressable pipeline. ### The Guardrails That Actually Matter Google is enhancing PMax transparency with channel-level reporting, asset diagnostics, search themes, and expanded negative keyword controls. These improvements allow B2B teams to differentiate branded demand from non-brand and diagnose lead-quality issues at the campaign level. However, these controls are only effective if configured properly. Here’s a quick guide: - **Brand exclusions:** Keep brand Search in a separate campaign. Don’t let PMax bid on branded queries and claim conversions that would have occurred anyway. - **Offline conversion tracking:** Feed SQL or pipeline stage data back into Google Ads. This is the most impactful change; without it, you’re optimizing to a proxy that may not correlate with revenue. - **Incrementality testing:** Run a holdout. Pause PMax in one geographic area or segment for two weeks and measure the actual lift in qualified pipeline. Platform-reported conversions are directional, not definitive. - **Budget guardrails:** Limit PMax to a percentage of total paid spend (start with 15-20%) and scale only after validating pipeline impact. Don’t let it absorb your Search budget by default. The hypothesis: adding offline conversion feedback and brand exclusions to PMax will materially reduce cost per SQL, although volume may initially decrease. This drop is expected. Your next step is to expand reach using the same qualification signal, not reverting to broad defaults. Success means keeping cost per SQL below your threshold. Guardrails ensure pipeline velocity doesn’t decline. A stop-loss should be implemented: if qualified pipeline drops more than 25% over four weeks, pause and diagnose. ### What the Numbers Actually Say Google claims PMax delivers 18% more conversions at similar CPA when added alongside existing Search campaigns. However, this claim isn’t specific to B2B SaaS and doesn’t address SQL or pipeline quality, potentially overstating value for teams focused on revenue outcomes rather than form fills. Benchmark data presents a clearer picture: PMax can yield dramatically cheaper leads, but cheaper leads don’t translate to cheaper pipeline unless your measurement system connects the two. The 104-account audit illustrates this well. The difference between $2,140 and $620 per SQL wasn’t due to campaign structure or creative refresh—it was the feedback loop. Teams that treat PMax as a supplement to Search, test with limited budgets, and evaluate incrementality against qualified stages will maximize its potential. Those who hand it the keys and take platform dashboards at face value will celebrate $25 CPLs while questioning why their pipeline remains flat. The pitch was always maximum performance with minimum effort. The real effort lies in measurement, not in the campaign itself.