One of the most common questions I get from B2B teams or anyone selling high-ticket products: “Meta just doesn’t get it. We’re not an e-com brand. Our deals close in weeks, not clicks. So how the hell do we train these campaigns?”
Yep. Been there.
I once ran a project where clients converted 2–3 weeks after filling out a form. Meanwhile, Meta was calmly sitting in Ads Manager waving the Learning Limited flag and pretending that everything was fine.
Spoiler: it wasn’t.
Meta loves fast, frequent, obvious signals. But when your product requires actual thought before purchase (scary, right?), the algorithm doesn’t know what to do with itself. Instead of learning, it throws spaghetti at the wall. And you pay for every noodle.
Here’s what usually goes wrong:
• You don’t get enough purchases to optimize around.
• The actual conversion happens 10–20 days later.
• Retargeting audiences are so small they expire before they start.
• Meta has no idea how to link a lead to a sale — because, well, she’s not that smart.
So campaigns just “run” — no learning, no optimization, just budget going quietly into the void.
Here’s what’s actually worked in long-sales-cycle projects:
1. Train Meta on the right steps in the funnel — not just the final sale. If your conversions take weeks, optimizing for purchases is like training a dog to sit by feeding it steak once a month.Start with earlier, more frequent events that actually mean something:
• form submissions.
• demo requests.
Meta doesn’t need magic. She needs repetition.
2. Use Conversions API with event_id to connect offline conversions to clicks. If someone buys 5 days after clicking your ad, you can still give Meta credit. Browser + server together = better attribution and actual learning. (Meta loves her data tidy.)
3. Send a custom “qualified lead” event based on CRM logic. If your sales team gives a lead the green light, you can feed that back to Meta as a signal. It’s not just filtering out junk — it’s telling the system: this is what good looks like.
And no, Meta won’t just figure it out. She’s not your strategist — she’s a pattern-matching engine. Give her the patterns.
So what actually works:
• Training on steps that happen before the purchase
• Sending delayed events with proper identifiers
• Giving Meta clear signals of who your actual customer is
• Accepting that you’re smarter than the algorithm — and acting accordingly
Meta can work for B2B. She can even work for long sales cycles. But only if you do the work to teach her how.
If you work with GA4 to BigQuery exports, be sure to check out my SQL cheat sheet.
