After a storm, the first callback gets the job

Here is the sequence, and it runs the same way every time. Wind or hail comes through on a Saturday. The homeowner walks out at 4pm, sees granules washed into the gutter run and 2 tabs lying in the driveway, and gets worried. The phone comes out. They search, they tap the first 3 map results, they tap a couple of ads, and inside about 20 minutes they have called 4 or 5 roofers.

Somebody answers. That company gets on the calendar for Monday. The rest get a voicemail box, and 2 of them never call back at all.

By Tuesday the homeowner has a contract and a claim number, and the roofers who lost it never find out they were in the running. They do not experience it as losing. They experience it as a quiet week after a storm, which they put down to the storm not being that bad.

What you can verify, and what you should ignore

There is a whole genre of speed to lead statistics circulating in contractor marketing, most of it traced back to studies nobody can produce on request. Ignore all of it. You do not need a study, because the numbers that matter are sitting in systems you already own and they are specific to you.

Here is what you can measure this week, for free:

• Your missed call count. Every phone system keeps a call log. So does a cell phone.
• Your real callback time, measured as the gap between a missed call and the first outbound call to that same number.
• How many missed calls never got a callback at all. This is usually the number that ends the conversation.
• How long your web form takes to reach a human, which you find out by filling it in yourself.

Everything past that is arithmetic you do with your own close rate and your own average job value.

Measuring your own response time

Pull 90 days of call detail. If you run a VoIP system (RingCentral, Dialpad, CallRail, Google Voice, any of them) there is an export. If the business runs off a cell phone, your carrier account has the same log, or you scroll it.

Then do this by hand, once. It is tedious and it is worth it.

1. Mark every inbound call that was not answered. Going to voicemail counts as not answered.
2. For each one, find the outbound call back to that number. Write down the gap in minutes. If there is no outbound call, write never.
3. Take the median, not the average. A single call you returned in 45 seconds does not offset 9 you returned the next day, and an average will let you believe it does.
4. Count the nevers separately, as a raw number and as a share of total missed.
5. Split out Saturdays, Sundays and anything after 5pm. That is when storm damage gets noticed, and that is where the never bucket lives.

Then test the form yourself. Take your own phone, off wifi and on cellular, and submit the contact form on your site with a real phone number. Time 3 things: how long until any automated acknowledgement arrives, how long until a human contacts you, and whether the lead landed anywhere you can find it afterwards. Web forms break silently. A form posting to an email address nobody reads anymore is one of the most common failures on a contractor site, and it throws no error at all.

Do the same for Google Business Profile messaging. If chat is switched on and nobody watches it, switch it off. An unanswered message is worse than no message channel, and Google shows your response time on the listing.

What to automate

The goal is not to replace a human. It is to close the 10 minute dead zone where the homeowner is still deciding and you are still up a ladder.

Missed call text back

An automatic text goes out within seconds of any missed call. For a roofer this is the highest value automation on the list, because it turns the exact moment you lost the call into a conversation you can still win.

The wording matters more than people think. A generic “sorry we missed you” gets ignored. Name the company, say when a human will call, and give 2 things they can do right now:

This is Dave at [company], sorry we missed you. If this is storm damage, text me your address and I will get you on the inspection list. Or book a time here: [link]

2 practical warnings before you switch it on:

• Register the number. Business texting in the US has to be registered for A2P 10DLC through your carrier or messaging provider. Unregistered traffic gets filtered, and you do not get a bounce notice, so you will believe the automation is working while the messages go nowhere. Check the registration status in your provider’s console.
• Test it from outside. Call your own main number from a phone that is not on your account, let it ring out, on a weekend, at the hour it will really fire.

A booking link that survives a thumb

If the text sends them to a booking page, that page has to work on a phone, in a driveway, operated with a thumb. Test it that way, not on your desktop.

• Under 30 seconds from tap to confirmed, or they quit.
• Ask for the minimum: name, mobile, address, time. Insurance details can wait for the inspection.
• Show real availability. A calendar offering slots you cannot honour is worse than a phone number.
• Confirm by text, not only by email.
• Load it on cellular data. A widget that appears in 2 seconds on office wifi can take 9 on a weak signal, and that is where it gets abandoned.

Form and after hours handling

Route web form submissions to a text on the phone of whoever is on call, not only to an inbox. Send an automatic acknowledgement that names a time (“someone will call you within the hour”), then make that time true. And run a different after hours greeting during storm weeks than the one you use in February.

Staffing the phone through a surge

Automation buys you minutes. It does not book roofs. Somebody still has to answer.

• Build the roster before the storm, not during it. Named wind events give you 48 to 72 hours of warning. Hail gives you almost none, which is exactly why the roster has to already exist.
• Cover 7am to 9pm for the 5 days after. Evening calls are people who got home from work and walked the yard in the last of the light.
• Take the owner off the roof for those days. The most expensive thing in a surge week is the one person who can sell an inspection spending the afternoon on a tear off.
• If you use an answering service, use one that books. A service that takes a message just moves the delay somewhere you cannot see it. It needs access to the real calendar and a script.
• Cut the intake script to 5 fields: address, mobile, what they saw, when they saw it, and whether another roofer has already been out. Insurance carrier is a 6th if the conversation allows it. Anything longer turns a booking call into an interview.
• Be honest about capacity on the calendar. If the crew can do 12 inspections a day, the booking tool should offer 12 and the rest get a real date. A homeowner given a Thursday slot on Saturday stays booked. A homeowner promised Monday who then gets nobody calls the next roofer.

Doing the arithmetic on your own numbers

Once the callback report exists, the value of fixing it is a multiplication you do yourself. Fill in your own figures instead of these:

• Inbound calls in your worst storm week: call it 40.
• The share you missed, from your log: call it 30 percent, so 12 calls.
• The share of those you never returned: call it half, so 6 people.
• Your close rate on a roof you actually get out and inspect: call it 1 in 4.
• Your average job value: whatever it truly is.

Take 6 people at a 1 in 4 close rate and you are looking at 1 or 2 jobs inside a single week. Run it with your real numbers and you get a figure you believe, which is the only kind worth acting on. And note what it is not: it is not an advertising problem. Those calls were already paid for.

The part worth remembering

A storm does not create demand so much as release it all at once, into a window that closes inside a day or 2. The roofers who do well in that window are usually not the best roofers in the county. They are the ones whose phone got answered.

All of this is cheap. The call log is free, the form test takes 5 minutes, text back automation costs less per month than a bundle of shingles, and the roster is a piece of paper. None of it improves your marketing. It stops you throwing away the marketing you already bought.

Written by Matt Pulliam, Southern Standard Systems. Wilmington, NC.