Search Partners on a small budget: switch it off or let it run?

Search Partners tends to get left on by default and then forgotten. Google pitches it as extra reach, but partner sites usually convert worse than Google's own results, and the placement data is so thin that it's hard to see where the waste is coming from. On a small budget where every click counts, I'd turn it off at campaign level and watch the numbers for a few weeks.

The counterargument is that some accounts do get decent volume from partners, and cutting it can starve a campaign that's still learning. A lot of it depends on whether you're spending a few hundred a month or several thousand. Has anyone run a clean split test with reliable conversion data, or is everyone going on gut feel?

Agreed on turning it off with a small budget, but the reasoning matters more than the split test. Partner traffic is mostly a volume issue: if a campaign can't spend its budget on Google's own results, partners just soak up whatever's left at a worse rate. So a simple rule works. If the campaign is limited by budget or impressions, switch partners off. If it's already beating its target CPA with room to spare, that's when it makes sense to test them on.

Your budget-limited test makes sense, though I'd check one thing before you flip it. Split the campaign by network segment and compare Google Search and Search partners as separate rows, looking at CPA over the last 30 days. If partners are clearly worse, that's the waste you're worried about, since they're only picking up the clicks Google's own results didn't take. If the gap is small, keeping them on is reasonable even on a tight budget, because the partner clicks aren't costing you much more per conversion. This also gives you a more direct answer than a split test, since you're reading your own data instead of guessing at the placement mix. The catch is that partner conversion volume can be small, so give it enough time before you trust a CPA gap.

Agreed on the split, but conversion lag can make partners look worse than they are. If your conversions often land days after the click, the most recent week of data is incomplete, and that hits partner rows harder when their clicks convert slower. Start with the conversion lag report for your action, then compare a 30-day window that ends far enough back to catch most of the late conversions. On a small budget, a few days of missing data can swing the CPA gap enough to flip your decision.

I'm not convinced partner clicks convert more slowly. Conversion lag mostly comes from the action itself. A lead form that ends in a sales call takes days to close whether the click came from Google's results or a partner site. Both segment rows are measured against the same conversion action, so the lag should hit them about equally unless your data says otherwise. Check the lag report split by network before you adjust for it, because the result might surprise you.

On a small budget the real problem is sample size. Partners might bring in 10 or 15 conversions a month, and a few missing days can swing CPA a lot. That's noise, not something specific to partners. A longer window, say 60 to 90 days if the account has the history, will do more for you than trying to model lag by network. I wouldn't call a CPA gap until both rows have enough conversions to be stable.

A longer window isn't the fix for lag. The end date is. Whatever window you pick, the last one to three weeks still hold clicks that haven't converted yet, so that stretch looks worse for both rows. A 90 day window ending yesterday has the same hole at the end, just with more history in front of it. Pull the end date back far enough that most late conversions have landed, then compare. You'll lose some recent data, but the CPA you get will be closer to what's real.