A canvassing route is not a driving route. That one sentence explains why most sales route-planning software is a poor fit for roofing, and why crews who buy it end up back on paper.
Route optimization exists to solve a B2B outside-sales problem: a rep has twelve scheduled accounts scattered across a metro and needs the shortest drive between them. Door knocking has the opposite shape. Every house on the street is a stop. The travel time between stops is eight seconds of walking. There is nothing to optimize.
What a storm crew actually needs from route planning is different, and simpler: which streets, in what order, and what's already been covered. This guide covers how to build that.
Step 1: Start from the damage, not the map
The single largest factor in a canvassing day's output isn't the route. It's whether the houses on it have damage.
Before drawing any boundary, pull the storm data for the area and answer three questions:
- Which streets were inside the swath, and what hail size did they take? A quarter-inch fringe and a two-inch core are completely different sales conversations.
- What's the date of loss? The date the storm hit is what the whole claim will be built on, and knowing it before you knock lets a rep open with a fact instead of a question.
- How fresh is it? The first week after a confirmed event is worth more than the following month. Freshness outranks convenience when you sequence.
A route built from a hail overlay looks different from one built by eye. It's often patchier — hail doesn't respect subdivision boundaries — and that patchiness is exactly the information you're paying for. Our guide to reading a hail map covers what the bands actually mean.
Step 2: Size the day to real doors
The most common territory-planning mistake is assigning an area far larger than anyone can work.
| Input | Realistic figure |
|---|---|
| Doors per hour | 15–25 |
| Productive canvassing block | 4–5 hours |
| Doors per day | 60–100 |
| Typical suburban streets to cover that | 6–10 |
Assign a rep ten times that and one thing happens every time: they cherry-pick. They work the easy streets, the nice houses, the blocks near where they parked, and the territory never actually gets finished. Then nobody knows what's covered, because "assigned" and "worked" have quietly become different things.
Size a day's territory to a day's doors. If a rep finishes early, hand them the adjacent block. That's a much better problem than a half-worked territory of unknown state.
Step 3: Sequence the streets
Within a day's territory, order matters. The sequence that works:
- Highest hail size first. Best damage, best conversations, best use of the rep's freshest energy.
- Freshest event next. If two areas took similar hail, work the newer one — fewer competitors have been there.
- Then geographic flow. Only after the first two are settled do you make the walking path sensible: down one side, back the other, no crossing back and forth.
- Save the marginal blocks for the end of the day. They're worth knocking, but not at 4pm when doors start opening.
The instinct to plan a tidy geographic loop first is strong and it's wrong. Tidiness saves a rep a few minutes of walking. Sequencing by damage changes what happens at the doors.
Step 4: Log every door as you go
A route plan is worthless without a coverage record, and the record has to be created at the door, not reconstructed later.
Every house gets a pin with an outcome: not home, not interested, inspection set, signed, come back Tuesday. That pin carries a GPS location and a timestamp, which turns coverage from a claim into a fact.
This is what makes the rest of it work:
- The saturation map shades what's been worked, so the whole team sees the same picture in real time and nobody parks on a street that's already done.
- The not-homes become the second-pass list, automatically, instead of being lost.
- The knock count becomes trustworthy, which is the only way a knock-to-contract rate means anything.
- The route plan self-corrects. Tomorrow's territory starts at the edge of what actually got covered today, not where the plan said it would end.
End-of-day tally sheets cannot do any of this, because the information arrives after the decisions that needed it.
Step 5: Plan the second pass before you need it
Here's the number that reframes route planning: on any single pass, 60–75% of doors don't open. Not-home is the majority outcome, not the exception.
Which means the second pass isn't cleanup. It's most of the opportunity, and it deserves as much planning as the first.
The rules that make a second pass work:
- Wait 2–4 days. Long enough that the household's schedule has moved, short enough that you're still ahead of the competition.
- Change the time of day. A street worked Tuesday at 2pm should be worked Thursday at 6pm. Same houses, largely different people home.
- Skip the hard no's. "Not interested" doesn't need a second knock. "Not home" does.
- Lead differently. "I stopped by Tuesday and missed you — I'm the one who's been inspecting roofs on this street" is a much warmer open than a cold introduction.
A crew that runs disciplined second passes will get materially more out of the same territory than a crew that treats a street as finished after one walk. Our post-hailstorm canvassing playbook goes deeper on the timing.
What route-planning software actually solves (and what it doesn't)
If you're evaluating tools, here's the honest split.
| Feature | Useful for roofing canvassing? |
|---|---|
| Multi-stop drive optimization | Rarely — every house on the street is a stop |
| Territory drawing and assignment | Yes — the core of the job |
| Live coverage / saturation shading | Yes — the thing that prevents double-knocks |
| Outcome logging at the door | Yes — the foundation everything else sits on |
| Storm / hail overlay on the same map | Yes, and it's the differentiator — it decides which streets are worth walking |
| Lead scoring from demographic data | Sometimes — much weaker signal than actual hail size |
| Check-in / mileage tracking | For payroll, not for planning |
Generic sales territory mapping tools are strong on the first three rows and have nothing for the fifth, which is the row that decides a storm crew's month. That gap — and what to look for instead — is the subject of our sales territory mapping software guide and the canvassing app comparison.
A worked example: Tuesday morning after a Sunday-night storm
- 7:30am — Pull the hail swath for the metro. Two-inch core over the northeast subdivisions, three-quarter-inch band running southwest. Date of loss: Sunday.
- 7:45am — Draw three territories over the core, each about eight streets. Assign one per rep.
- 8:00am — Reps see their boundary on their phone with the hail layer under it. No printed maps, no group text explaining where to go.
- All day — Every door gets a pin. The coverage shading fills in live. The manager can see at 11am that one territory is running slow and move a rep.
- 4:00pm — The last two hours go to the streets with the most doors, because that's when people are home.
- Wednesday — New territories start at the edge of what's shaded, not at a guess. The Sunday-storm core gets finished before anyone touches the three-quarter-inch band.
- Thursday evening — Second pass on Tuesday's not-homes, at a different hour.
None of that requires optimization math. It requires a map that knows where the hail was and a record of where people have already been.
The bottom line
Route planning for door knocking is three decisions: which streets have damage, how many doors is a real day, and what's already been covered. Get those right and the walking path sorts itself out. Get them wrong and no amount of optimization saves the day.
The piece most crews are missing isn't a route planner — it's the storm layer underneath it. HailMate puts the hail swaths, the territory boundaries, and every logged knock on the same map, so the route plan and the damage data are the same object instead of two things a manager reconciles at 7am.