The cheap leads came before me. I watched the funnel for four months: they filled the CRM and rarely became qualified.
Decision
So I changed the target to leads that moved from MQL to SQL, and cost per qualified lead and cost per sale became the measures. The measurement was built first, so the case rested on what the leads did further down.
Who I had to convince
The people who approve the budget. It's an awkward case to make: pay more for each lead, and wait.
What changed
Those two are the measures we steer by, and bidding now learns from qualified pipeline.
How I'd have known I was wrong
If cost per sale hadn't moved within two quarters, I'd have made marketing smaller rather than sharper, and I'd have said so.
