Conversion count setting: one per click or every conversion for lead gen?

The usual advice is "one" for leads and "every" for purchases, and I'm not sure it holds for every form. Take someone who submits a lead form, then comes back a day later through the same click and submits again. Sales could fairly call that two real leads, but the platform only saw one click. Counting every conversion there inflates your numbers and can push Smart Bidding toward the wrong signal. My instinct is one per click for lead gen, and every only where a repeat visit means a repeat sale. The honest answer probably depends on how often the same person submits twice, so it's worth checking against CRM data. Has anyone actually done that comparison?

I agree with the instinct, but I'd say the deciding factor is whether a second submission creates new sales work. In most CRMs, a repeat fill from the same email about the same thing within a day ends up as one opportunity anyway, so counting "every" would just inflate the numbers. A repeat that's a genuinely new request, like a different location or service, is a separate lead and deserves its own count.

The setting also only governs one click. If the same person comes back through a fresh ad click, Google records a new click and can count it under either setting. So a gap between Ads and the CRM usually comes down to how duplicates are handled upstream, not the count setting itself.

Right on the new-click part, and that's why the setting matters less than it looks. "One per click" only collapses repeats that share the same gclid. Someone who comes back next week gets a fresh click, so both settings count the second lead, and the setting can't stop it.

The dedupe has to happen where the lead is created. If the form writes a lead ID, or the CRM matches on email plus service, you can skip the second conversion, or send it as an offline import tied to the original lead so it updates the first one instead of adding a new one. That way Smart Bidding learns from one lead, not two.

Take a dental clinic. A patient clicks a brand ad, submits a form, then clicks again three days later and submits for the same cleaning. That second one should be suppressed. If the same patient comes back asking about implants, it's a separate lead and counting it is fine.

Simplest test I know, and it doesn't need CRM data each time: would sales treat the second submission as new work? If it's the same person asking about the same service, it's one lead no matter how many days apart. A different service or location counts as a new lead.

The rule holds because Smart Bidding only needs a signal that tracks real sales work. If you count the duplicate, you teach it that this traffic produces more leads than it actually does, and it will bid to chase that.

"No matter how many days apart" is too broad. If the same person asks about the same service a month later, after the first lead went cold, that's often real new work. A rep has to re-qualify them, and plenty of CRMs reopen the old record as a fresh opportunity. A repeat within a day or two is clearly one lead, but past that it depends on how your CRM handles old leads. If you suppress every repeat regardless of timing, you undercount the people who come back and buy, and Smart Bidding ends up learning from a signal that's too low.

Agreed, and the cutoff is the part most people skip. Before you settle on a window, look at how your CRM ages leads. If it auto-closes a lead after 14 or 30 days, anyone who comes back after that is a new opportunity, so count it. Inside that window, suppress it.

Makes sense, but where you draw the cutoff really depends on whether your CRM reopens the old record or creates a new one when the same email comes back. If it reopens, the repeat is real re-qualification work even inside your window, so count it. If it creates a new record, the second submission is usually a duplicate, and suppressing it is the right call.

I wouldn't drop the suppressed repeats from measurement entirely, though. Log them as a secondary action, or send them as an offline update tied to the original lead, so you can see how many there are each week. If that number is high, the problem is probably the form or the landing page sending people back, not the count setting.