Offline conversion imports

The central problem in leadgen is that the ad platforms optimise toward the conversion they can see - the form fill - and the form fill isn’t the outcome you care about. Left to default, the platform’s algorithm learns to find people who fill in forms, which is not the same population as people who become customers, and often actively the opposite. You get cheaper leads and worse ones.

Offline conversion imports fix this at the source. You send the eventual sales outcome - this lead became an opportunity, this one closed for £8,000 - back to the platform, matched to the original click. The platform’s bidding then optimises toward the leads that actually closed rather than the ones that merely converted on-site. It’s the single most effective way to align paid acquisition with lead quality instead of lead volume, and it’s underused because it’s plumbing, not a campaign setting.

How it works

The mechanism is the same idea across platforms, under different names - Google’s offline conversion import, Meta’s conversions API with offline events:

1. Capture a click identifier when the lead arrives, stored against the lead in the CRM.
2. As the lead moves through the funnel cascade, the CRM records the stage reached and the value.
3. Push those outcomes back to the platform, keyed on the click identifier, so it can attribute the real result to the original ad.

This is a data-pipeline job as much as a marketing one, and it depends on the click identifier surviving all the way from ad to CRM - exactly the kind of identity plumbing that quietly breaks and nobody notices until the optimisation degrades.

Why this is the leadgen edge

It’s the answer to MQL theatre. Instead of arguing with the sales team about lead quality, you let the closed-won data retrain the acquisition engine automatically. The platform stops buying form-fillers and starts buying customers, and the whole cascade improves without touching the landing page. It’s the same instinct as DTC feeding real outcomes back rather than trusting on-platform conversions, applied to a funnel where the outcome arrives months later.

Where it breaks

• Signal volume. The platforms need enough conversions to optimise, and closed deals are rare and slow. A low-volume B2B funnel may not generate enough closed-won events to train the algorithm, so teams import an earlier proxy - a qualified opportunity - as a compromise between signal and volume.
• Match rates. If the click identifier is lost on the way to the CRM, the outcome can’t be matched back and the import silently covers only a fraction of leads. The optimisation is only as good as the match rate.
• Lag. A deal that closes in 90 days teaches the platform about an ad it ran a quarter ago. In fast-moving auctions that lag dilutes the signal, which is another reason an earlier-stage proxy is often the practical choice.