You can finally see what the Search Partner Network is spending

Search Partner Network spend is now visible in every account. Segmentation rolled out to all advertisers in January 2026, and separately an alpha lets some advertisers switch partners and Display off independently in Performance Max, which is the first real network-level control PMax has ever had.

For years the honest answer to what the Search Partner Network is costing us was a shrug and a guess. That answer is no longer acceptable, and neither is the reflexive decision that used to substitute for it.

What the network is, and why it was murky

The Search Partner Network is Google search results appearing on properties Google does not own. Site search on retailer and publisher domains, some app inventory, various search boxes across the web. It has always been bundled into Search campaigns by default, and until now the reporting rolled it into a single blended number with Google Search itself.

That created a predictable split in the market. One group left it on and never looked. Another group turned it off everywhere as a matter of policy. Both were guessing, and one of them was quietly leaving volume on the table.

The report to run this week

Open your Search campaigns. Segment by network, with search partners shown separately. Pull at least 90 days so you have enough conversions to say something. Then look at four things.

Share of spend. If partners are under 5% of spend, this is a rounding error and you can stop here. If they are over 15%, it deserves a real decision.

Cost per acquisition or return versus Google Search. The comparison you want is not is it as good, because it rarely is. It is is it good enough to be worth the volume.

Conversion rate. A large gap in conversion rate with a similar click cost points to intent quality, which is the usual failure. A similar conversion rate with a higher click cost points to auction dynamics, which is a different problem.

Time to conversion. Partner traffic often converts on a different lag. If your lookback is short, you may be crediting partners with less than they earn.

The decision rule

Do not switch off reflexively. That is the point of having the data.

A workable rule: if partner CPA sits within roughly 20% of your Google Search CPA, keep it on. The volume is real and the efficiency loss is inside the noise band of everything else you do. If partner CPA is double, and partners are a meaningful share of spend, turn it off and reinvest the budget in the campaigns that were budget-constrained anyway.

Between those two, look at the mix. If partner spend is concentrated in a handful of campaigns, treat it as a campaign-level decision rather than an account-level one. Turning it off account-wide because one campaign is wasteful is the same error as negating a whole cluster because one query failed.

The PMax angle

Until this alpha, the network question in Performance Max was all or nothing. You took the whole bundle or you did not run the campaign. Advertisers who wanted Search-like control with feed automation had no way to remove the parts they did not want.

Being able to switch partners and Display off independently changes what PMax is for. It becomes a configurable campaign rather than a package, which makes it viable for businesses that had a legitimate reason to avoid it, particularly regulated industries and brands with strict placement policies.

The alpha is not universal. If you have a Google representative, ask directly. If you do not, the segmentation report still gives you the decision data, and you can act on it in Search campaigns today while waiting for the PMax control to reach general availability.

What is still not visible

Partner placements are still not fully named. You get network-level performance, not a site list, so you cannot build an approved-placements whitelist the way you can in Display. For a brand with genuine placement sensitivity, that gap matters, and it is the one honest reason to remain cautious.

The mitigation is the account-level placement exclusion list, which does persist across campaigns, plus content and inventory type settings. Those cover the categories, not the specific properties.

What to do with the finding

Whatever the data says, write it down with a date, the spend share, and the CPA gap. Then set a calendar reminder to re-run it in six months.

The reason is that this is not a stable measurement. Partner inventory composition changes, and a decision that was correct in March is not automatically correct in September. Accounts carry an enormous amount of frozen judgement, made once on evidence that expired years ago, and network settings are one of the most common places it hides.

Run the segment, make the call, date the file, and you have converted a permanent argument into a scheduled review.