Your best channel for leads is probably your worst channel for customers.

Almost every lead-gen business measures form submissions, because that is what the website produces and it is trivially countable. It is also the single clearest example in web measurement of optimising the wrong number.

The mechanism is simple. Volume-oriented channels drive traffic with weak intent. Weak-intent traffic still fills in short forms — often at a higher rate than high-intent traffic, because high-intent buyers read more before enquiring. So the channel producing the worst leads posts the best conversion rate, wins the budget argument, and produces more of the same.

Meanwhile sales knows. There is always a salesperson who can tell you which source produces enquiries worth calling, and that knowledge lives entirely in their head because no report has ever agreed with it.

The conversion event has to be the qualified submission. Not the form fill — the enquiry that turned out to be real.

Every form counts the same. The gated whitepaper, the newsletter box, the contact form and the “request a quote” form usually fire the same event. Once pooled, a channel driving whitepaper downloads outranks a channel driving quote requests, because there are more of them. This alone invalidates most lead-gen channel reports.

Qualification lives in a different system. The submission is measured on the website; whether it was any good is decided in a CRM, days or weeks later, by a person. Nothing joins the two, so the outcome never travels back to the source that produced it.

Volume targets outlive their usefulness. Once a team is measured on lead count, quality problems become somebody else's. The measurement system built to serve that target then makes the misalignment invisible, because it has no column for the thing sales cares about.

Deployment here is unusual in one respect: the interesting work is at the join between the website and whatever decides that a lead was real.

• Each form treated as a distinct event with a distinct commercial weight, rather than pooled into one submission count.
• A qualified-lead definition agreed with sales, written down, and used as the ranking axis.
• A mechanism for the qualification outcome to travel back to the originating source, without us taking access to your CRM.
• Source, campaign and entry-page ranking against qualified leads, with raw submissions kept visible alongside so the gap between them is explicit.
• Behavioural scoring on the sessions preceding each enquiry, used as an immediate proxy while the real outcome is still pending.

That fourth point matters more than it looks. The gap between a source's raw submissions and its qualified leads is the most useful single number a lead-gen business can have, and almost nobody has it.

Source ranking against qualified leads, with the raw submission count next to it. Two columns, deliberately adjacent, because the divergence between them is the finding and burying it in separate reports is how it stays unnoticed.

Below that, quality rate by source over time — which surfaces the channel that was producing good enquiries and has quietly stopped, usually after a targeting change nobody told marketing about.

Then the visitor list ranked by conversion likelihood, which for a lead-gen business doubles as a call prioritisation order.

Measuring qualified leads instead of raw submissions costs you speed. A submission is known instantly; a qualification is known in days or weeks. Any team used to checking lead count daily will find the new number frustrating at first.

The compensation is the behavioural score, which is available immediately and correlates well enough with eventual quality to be worth watching between qualification cycles. It is a proxy and we label it as one. It is a considerably better proxy than raw volume.