Every roofer misses calls during a thaw, so decide now which ones you miss
When the melt arrives your phones will overflow, and deliberate triage beats a promise to answer everything.
A warm stretch in February releases every failed seam in the city at once. Call volume can double for two or three days and then drop back. Nobody staffs for that peak all winter, and pretending otherwise just produces voicemails nobody returns.
Plan the overflow before it happens. An overflow number that reaches a second person. A voicemail that states a realistic callback window instead of a cheerful promise. A text auto-reply carrying one short form.
Rank the calls in advance. Water coming in now goes first. Multi-unit buildings with tenants go next. Planned estimate requests can wait a day, as long as somebody tells them so.
Then measure the unglamorous number: how many inbound calls went unanswered last week, and what became of them. Most roofing companies have never looked once.
Sort the caller by building before you sort them by budget
Bungalow, two-flat, greystone or courtyard building tells you more in three seconds than any qualifying script.
Building type predicts most of what you need. Pitch or low-slope, membrane or shingle, rough square count, access, whether a tenant has to be coordinated with, and how the decision actually gets made.
Put the question near the top of the intake script, right after the address and whether water is coming in now. Owners answer it without hesitating, because it is how they describe their own house.
The rest of the call gets shorter and better. You know which crew, roughly which price band, and which photographs to send afterward.
Recording calls can be genuinely useful for improving the script, and consent rules vary, so it is worth confirming the requirements with your own counsel before you turn anything on. Whatever you do, listen to a handful each month. The question that ends sales is usually the same question.
You cannot inspect a roof under six inches of snow, so sell the next step instead
A winter appointment that cannot be completed is a leak in your pipeline, not a scheduling inconvenience.
From December into March a share of inspections simply cannot go ahead. Snow cover, ice, unsafe access. The crew leaves, the homeowner receives nothing, and the file goes cold within a week.
Give the estimator something to hand over anyway. An interior and attic look. Photographs from the ground and the eaves. An honest note on what can be judged now and what cannot. And a booked return date once the roof is clear.
A dated follow-up with a name attached converts far better than a promise to call when the weather breaks. Put it in the system as a task rather than in somebody's head.
Say plainly what is safe to walk and what is not. Refusing to guess about a roof you could not properly examine reads as competence, and it is also the right answer.
Proposals sent in November are not dead, they are waiting for April
In a freeze-thaw market a large share of lost proposals are deferrals with no follow-up attached to them.
A homeowner who receives a bid in late autumn often decides to get through one more winter. Nothing was wrong with the proposal. Timing killed it, and everybody involved knows it.
Build a spring reactivation list on purpose. Everyone who received a proposal after October and did not sign gets a specific message in early spring covering current pricing, crew availability, and what another winter of freeze and thaw generally does to the failure they already have.
Stay neutral where insurance is involved. Describe what you saw and what you recommend. Do not tell an owner what their carrier will decide, and point them to their own insurer when a claim is in play.
Follow-up beats discounting here. The buyer already wanted the roof. They wanted it in April.
Booked crew days is the number that pays payroll
Leads, appointments and close rates all move around; the number that decides the year is how many days each crew is scheduled.
A roofing company can improve every funnel metric and still have a crew sitting idle on a Tuesday in July. Track filled crew days per week, by crew, against capacity.
Work backward from there. How many signed contracts fill a week. How many completed inspections produce a signature. How many answered calls produce a completed inspection. Four numbers, reviewed monthly, are enough.
Split by geography once and only once: the city and near suburbs against the collar counties. Travel time and staging make them different businesses with different margins.
Attach a source to every signed job. Without it, every marketing decision next season is a guess wearing a spreadsheet.
A February number and a May number are not a test result
Seasonality here is larger than any change you will make, so compare like periods or do not compare at all.
One metro does not produce enough traffic for small-effect split tests on a roofing site. The honest options are large changes measured across long windows, and process changes you can observe directly.
When you do compare, compare against the same weeks last year or run two variants simultaneously. A conversion rate that doubled between February and May is telling you that spring arrived.
Test the things with big effects: the intake script, what the form asks, whether the estimator books the return visit on the roof rather than back at the office, whether proposals get presented or emailed.
Five recorded calls and ten follow-up sequences will teach you more this quarter than any button experiment. Keep a written log of what changed and when, or by next winter nobody will remember what worked.
Questions we actually get
- We answer every call. Is intake really our problem?
- Pull the phone records for the last thaw week and count unanswered inbound calls, then count how many were called back within the day. Nearly every roofing company that believes it answers everything discovers a cluster of missed calls during the two or three days each winter when the phones spike hardest.
- Should we record calls?
- Recording is one of the most useful sources of insight available to a contractor, and consent requirements vary by state and situation. Confirm what applies with your own counsel before switching anything on, use clear notice, and once it is running, listen to a small sample every month rather than storing recordings nobody plays.
- Do we need a CRM, or will a spreadsheet do?
- A spreadsheet works until proposals need dated follow-up tasks, which is exactly the point where roofing money leaks. Whatever tool you use must hold the address, the building type, the source, the proposal date and a scheduled next action with a name on it. The tool matters less than whether anyone maintains it.
- How do we answer insurance questions without overpromising?
- Describe what you observed and what the repair involves. Do not predict coverage, do not characterize what an adjuster will decide, and tell the homeowner to speak with their own insurer. A straight answer that stays in your lane is more persuasive than a confident guess that unravels three weeks later.
- Is A/B testing worth doing on our site?
- For headline and button variations, generally not. A single metro rarely delivers the traffic needed to resolve a small difference, and the seasonal swing here overwhelms it anyway. Larger structural changes measured over long windows are worth doing, and so is testing the parts of the process you can watch directly, such as the phone script and the follow-up sequence.
