Offline SQL uploads for Smart Bidding: switch the goal now or wait for volume?

Picture running Google Ads for a B2B services firm with a three-month sales cycle. Smart Bidding sees the form fill as the conversion, but sales only cares about the fills that turn into opportunities, maybe one in five. Uploading offline conversions for qualified leads looks like the obvious fix, but the bidder gets far fewer signals, so the campaign can sit in learning for weeks. My take: keep the form fill as the bidding goal until you're getting roughly 30 to 50 qualified conversions a month, then move the goal to the SQL stage and keep fills as a secondary column. Has anyone seen the opposite work at deal sizes over a few thousand?

Agreed on the threshold, but I'd check the lag before the volume. Google's offline conversion imports have a click-to-upload window, as far as I know 90 days from the click (confirm in the current docs). With a three-month cycle, a lot of SQLs from later clicks will fall outside it, so a deal that closes at day 100 may never reach the bidder at all.

Pull the median and 90th-percentile days from form fill to SQL out of your CRM before you pick a switch date. If the 90th percentile is near or past 90 days, the SQL goal will lag no matter how many leads you have, and you may need an earlier stage as the uploaded event, like a booked discovery call.

And if you do upload SQLs, send them with a pipeline value instead of a flat count. That gives Smart Bidding something to chase at the deal sizes you mentioned, rather than treating a small retainer and a big contract the same.

Checking the lag first is the right call, and I'd turn it into a simple rule: upload the earliest stage whose 90th-percentile lag stays under about half the window, so around 45 days if the window is 90. A late SQL doesn't just arrive late. If its click is older than the window, the conversion gets dropped and nobody sees it happen, so the bidder looks like it's underperforming when the data never reached it. Leave some headroom, because lag tends to drift up when sales is busy. And keep form fills in the client report so volume stays visible while the SQL goal does its slower work.

Half the window is a sensible rule, but I'd look at match rate before trusting any lag numbers. An offline upload only counts if the record carries a click ID, and B2B CRMs often lose that field when a lead moves from marketing to sales, or when an import overwrites it. If a third of your SQLs can't be matched, the bidder is learning from a smaller, skewed sample, and no amount of lag tuning fixes that. Pull the match rate by week before you flip the goal. What rate are you seeing on the SQLs you already have?