How do we tell a dead lead from one that is just slow?
Build the conversion curve from your own history instead of guessing. Take leads from a period old enough to have fully matured, then plot what share of them converted at each week of age. The point where that curve flattens is where slow becomes dead for your business. Segment it by job type, because an emergency repair is dead within days while a system replacement or commercial bid can still close months later.
The curve is easy to build and almost nobody has one
Pull every lead created in a window that ended long enough ago that essentially all of them have resolved. For each lead, record the number of days between creation and the first booked job. Then compute the cumulative conversion rate at day seven, fourteen, thirty, sixty, ninety and beyond.
The result is a curve that rises steeply and then bends. The bend is the answer to the question. Everything after the bend is the tail, and the tail tells you what late follow-up is actually worth.
Segment before you conclude anything
A single blended curve averages together customers whose water heater is leaking with customers pricing a replacement system for next spring. The blended bend is meaningless for both.
Split by job type at minimum, and by estimate value if you have it. Urgent repair curves are nearly vertical and then flat. Large-ticket replacement curves have a long, real tail. Commercial and multi-bid work has the longest tail of all, and killing those leads on a thirty-day rule throws away revenue.
The mistake that makes late conversion invisible
If you include recent leads in the calculation, you are counting leads that have not had time to convert as if they failed to convert. That censoring pulls the curve down and makes the tail look smaller than it is.
Only include cohorts old enough to be complete. If your longest realistic sales cycle is six months, your most recent usable cohort is six months old. This is the same maturation problem that distorts revenue reporting when people compare a fresh month against a settled one.
What to do with each region of the curve
Before the bend, aggressive follow-up is worth real labor: calls, texts, a person. After the bend, the economics change and the right answer is usually low-cost nurture rather than outbound effort, then a reactivation attempt at the point the equipment or need is likely to recur.
Wiring that transition automatically, so leads move from an active cadence into a long-cycle list without anyone remembering to do it, is exactly the kind of small operational decision that operational AI is good at, and it depends on the CRM knowing lead age and job type reliably. If that data is inconsistent, fix the customer data first.
Topics: lead lifecycle · follow-up · diagnostics · conversion
Have a version of this question about your own business?
The useful answer usually depends on which systems you run and how they're connected. That's a conversation, not a blog post.