Hard-Won Lesson
We Upgraded to Higher-Quality Leads. I’m Still Not Sure It Was Right.
In 2024 we rebuilt the door-to-door channel I’d run since 2019. The old model: team leads driving van crews to manage territory and generate volume. The new model: individual reps in branded vehicles running homeowner inspections, then flipping them to the sales team. Higher-quality leads, every conversation warmer, every appointment more likely to close.
Two years in, here’s a sentence you will not find in a case study: I’m still not sure it was the right call. This post is the evidence for both verdicts.
The Trade, In Numbers — Door Channel Before and After the Switch
What Improved
the channel’s close rate, 2023 → 2026 YTD — climbing every single year since the change
What It Cost
annual demos (2024) vs. 2026 year-to-date — plus higher pay per rep, on fewer leads
The case for the switch
Every quality signal moved the right way, immediately and durably. A homeowner who’s had a professional inspection of their actual roof is a fundamentally different appointment than one who agreed to a pitch at the door — the close rate’s four-year climb says the sales team agrees. The branded individual vehicles put more wrapped trucks on more streets, which feeds everything we’ve written about job-site fame. And the model contributed to a company-wide funnel transformation we’ve published with receipts. By every per-lead measure, the new model produces a better lead.
The case against it
Per-lead quality is not channel economics. The volume floor fell out — 879 demos became 588 became 91 year-to-date, and a ten-point close-rate gain doesn’t offset an order-of-magnitude volume loss. The reps cost more individually: an inspector with a branded vehicle is a heavier unit of payroll than a canvasser in a shared van, paid on a thinner stream of opportunities. And the channel’s cost of marketing got rougher after the change — my own assessment from inside it, with the full fully-loaded spend detail being exactly the ledger work we’re closing next. The old model produced a documented $1.99M year at 15.5%. The new one hasn’t yet produced a ledger that brags like that, and honesty means saying so.
Quality-versus-volume trades look obvious on a slide and ambiguous in a ledger. Better leads at lower volume and higher labor cost is three variables, and the slide only shows one.
The three numbers that will settle it
Undecided doesn’t mean unmeasurable. The verdict comes down to: one — the model’s fully loaded cost of marketing, rep pay and vehicles included, against the 10–15% band the old model lived in; two — the volume floor: whether inspections can be recruited and scheduled back to a demo count that moves company revenue, or whether the model structurally caps out; three — blended contribution: total channel-attributed sales against what the same payroll would produce under the volume model. When those three are on one page, the question answers itself. Until then, anyone declaring victory in either direction is publishing a preference, not a result.
Why publish the undecided version at all? Because this exact trade — fewer, better, more expensive — is sitting on the desk of half the contractors we talk to, dressed up as obvious by whoever’s selling the new model. It isn’t obvious. It’s a three-variable bet that took us two years to even instrument properly. Make it with the instrumentation, not the slide.
We upgraded the lead. The jury’s still out on the channel. Both of those are facts, and only one of them fits in a case study.
Facing a Quality-vs-Volume Decision?
The free audit instruments the trade before you make it — fully loaded cost, volume floor, and blended contribution, on one page.
