Do restoration companies miss more calls than other home service businesses?

A water job at 2am feels like the call most likely to slip through. The phone's in the truck, the tech on call is already standing in somebody else's basement, and the office went dark at five. So it's a fair question whether restoration shops drop more calls than plumbers, HVAC outfits and the rest of home services. There's a published benchmark that answers it, and the answer is less dramatic than the 2am picture.

Here's the short answer. Not by the benchmark. In Invoca's 2026 home services report, a person answered 51% of calls to restoration companies, against 52% across home services. That puts restoration seventh of the nine groups Invoca measured, two points behind doors and windows at 53% and well ahead of HVAC at 34%. It's an average across Invoca's own customers, and it can't tell you your shop's rate. You get that by dividing two numbers off your own phone records.

Full disclosure: Willison sells an AI receptionist, so we have a side in any question about unanswered calls. The method below works whoever answers your phone, and the parts about Willison say so.

The math in one line

Answer rate is the calls a person picked up, divided by every call that came in. Invoca's 2026 home services benchmarks ran that division across the calls its platform tracks for restoration customers and got 51 out of every 100. Across all nine home services groups, it got 52.

The other 49 aren't 49 lost jobs. The denominator is every tracked call, so it holds misdials, robocalls, and people who hung up after one ring. Invoca shows how much that matters across home services as a whole: count only calls longer than 15 seconds and the answer rate climbs to 65%. Over 30 seconds, it's 73%. Invoca doesn't publish those filtered rates by trade.

Don't read the filtered number as the true one, either. A homeowner with water on the floor who hears four rings and hangs up to try the next name on the search page was on the line for under 15 seconds too. A short call can be junk, or it can be the job you most wanted. The honest picture sits somewhere between the raw rate and the filtered one.

Where restoration sits among the nine groups

Here's the full answer-rate list from Invoca's 2026 home services report, highest first. Not every group on it is a trade we write about; it's here so you can see the whole spread.

1. Plumbing: 74%
2. Pest control: 73%
3. Carpeting: 65%
4. Home security: 57%
5. Lawn and trees: 54%
6. Doors and windows: 53%
7. Restoration: 51%
8. HVAC: 34%
9. Construction: 32%

Restoration sits one point under the home services figure. The gap between the top and the bottom of the list is 42 points. Next to that, restoration's distance from the home services figure is noise.

The bigger story on the list is the 40-point gap between plumbing and HVAC. From outside, those two phones look a lot alike: emergencies, weather swings, calls landing while every tech is on a job. Invoca records the outcome, not the cause, and its customer mix may not be the same from trade to trade, so the list can't tell you why plumbing answers more. What it does show is that "we're an emergency trade" doesn't settle the number in either direction.

What 51% can't tell a restoration owner

It's an average, so the worst nights are diluted. Invoca says its figures are averages across its customer base and doesn't say what period they cover. If that period includes a winter, its freeze nights are blended in with the quiet Tuesdays. Water losses don't arrive evenly. They bunch up on cold snaps and storms, so your rate on the worst night of the year can sit a long way from your rate across the year.

It doesn't say which calls went unanswered. On a restoration line, the call nobody picked up might be a misdial, a supplier, or a homeowner whose basement is filling. Those aren't worth the same. The EPA's guide to mold and moisture says water-damaged areas should be dried within 24 to 48 hours to prevent mold growth, which is why a water call doesn't keep until Monday the way a quote request might.

It only sees the phone. The benchmark covers calls on the numbers Invoca tracks. Some restoration work arrives by other routes: a program assignment by portal or email, a plumber who sends you jobs, an adjuster or a property manager ringing the owner's cell. An answer rate measures what's riding on your direct calls, not your whole pipeline.

It's Invoca's customers, not every shop. These are businesses that pay for call tracking. Nothing in the data says whether that makes them better or worse at answering than a shop that doesn't, so use 51% as a reference point, not a verdict.

Worked example: one month of your own calls

The benchmark gives you something to hold your number against. The number that matters is yours, and you can get it in an afternoon. The catch is that a carrier's call log has no column called "answered by a person." It shows whether a call connected and how long it ran, and a call that reached voicemail can show as connected.

1. Pull every inbound call for one ordinary month from your carrier or phone system, with the time and duration of each.
2. Merge in the after-hours phone before anything else. If nights forward to an on-call cell, the main line may show those calls as forwarded, as missed, or not at all. Pull that cell's history for the same dates, keep only the inbound calls that came through the forward, and match each one by time to its line in the main log so every call counts once. Skip this and your night numbers will look worse than they are.
3. Take out the calls that went to voicemail. If your phone system records which device or extension picked up, use that. If it doesn't, match each message in your voicemail box to its line in the log by time, and mark it unanswered. Then time your greeting: a connected call with no message that ended within a few seconds of that length hung up at the greeting, so mark it unanswered too.
4. If none of that comes out clean, tag the next seven days as they happen: whoever answers marks each call as answered, voicemail, or rang out. A week you trust beats a month you don't.
5. Divide the calls a person answered by the total inbound.

Say the month shows 180 inbound calls and, with voicemail taken out, 99 reached a person. That's 99 divided by 180, or 55%. Those are placeholder numbers to show the arithmetic. They aren't a benchmark and they aren't a guess at your shop.

Before you hold 55% against 51%, keep the comparison loose. Invoca doesn't say which lines its customers track, so its mix of callers may not match yours. Your log also holds adjusters, suppliers, subs, and customers asking where their job stands, and those get answered at their own rate. To see how you do with people who don't know you yet, take out the numbers already in your job or customer list and run it again. None of these cuts will line up exactly with Invoca's, which is one more reason to treat the benchmark as a reference point and not a scorecard.

Then run it a second way, counting only calls over 15 seconds, and hold that against the 65% figure for all of home services, not against 51%. Filtered goes with filtered.

Second worked example: the same month, split by the clock

One monthly rate can hide a weak slice. Split the same month into business hours and everything else, and divide each one on its own. Your monthly rate is just those two rates averaged, weighted by how many calls each slice holds.

Take the same 99 answered out of 180. Say 120 of the calls came in during business hours and 60 at night and on weekends. The 99 could be 72 of 120 by day and 27 of 60 after hours, which is 60% and 45%. Or it could be 66 of 120 and 33 of 60, which is 55% and 55%. Same month, same total, two very different phones.

If your two slices land close together, your monthly number is honest. If they're far apart, it's hiding the slice that matters, and for a water shop the slice that matters is whichever one the losses arrive in.

Then do one more: your worst weather week from last season, on its own. Label it as what it is. It's the week your phone carried the most, not a normal week, and it shouldn't stand in for one. It's still the week you most need to see.

Where owners get the number wrong

Some mistakes make your rate look worse than it is:

• Leaving out the on-call cell. Calls a tech picked up at 1am were answered by a person. Merge them in.
• Holding your raw rate against a filtered benchmark. 65% is the over-15-seconds figure. Your raw count belongs next to 51%.
• Treating a storm week as typical. It's the hardest week, so it belongs in its own row.

Others make it look better than it is:

• Counting voicemail as answered. It connected. Nobody picked up.
• Holding your filtered rate against the raw 51%. Same mistake as above, pointed the other way.
• Counting business hours only, because that's the log the office sees. It flatters the rate whenever your days run better than your nights.

And one belongs in neither column: callbacks. Track how many unanswered calls got a call back, and how fast, as its own number next to the answer rate, not folded into it. A quick callback beats no callback, but on an active loss the homeowner may have booked someone else in the meantime, so it isn't the same thing as an answer.

What changes the unanswered half, and what each option costs

Strictly, only the first three options below raise the share of calls a person answers. The fourth changes who answers. Set it to catch only the calls nobody picks up, after hours and on missed calls, and it adds to what you have. Send every call to it, and you trade a person who can commit a crew on the call for an answer every time.

• You and your techs answer more. No new bill. It costs attention on the job and sleep at night, plus on-call pay where you pay it. The upside is real: the caller reaches someone who knows the work and can commit a crew and a time on the call, which nothing else on this list does. It fails the same way every time: a person answers one call at a time, and the one in a crawlspace answers none.
• Forward nights to an on-call cell. Cheap to set up, and a person who knows the work still answers. It keeps that upside and inherits every limit above, and it can drop calls out of your main line's log, so the count has to be merged by hand.
• A human answering service. A trained person on the line, at any hour. Depending on the service and the script you give it, it takes the details or patches the call through to your on-call phone. It costs a monthly bill, priced differently by different services. It fails when the script doesn't ask what you need to know, so your tech calls back cold, or when a patch rings out.
• Willison. Willison answers in seconds, at any hour, and answers every call at once, so the second caller on a freeze night is answered as fast as the first. It asks where the water came from, whether it's still running and how long it's been going, takes the address, and hands those details to your team so your tech can judge the category before rolling. It doesn't dispatch crews or promise an ETA. It costs a monthly subscription, with prices on the pricing page. It can mishear an address or misclassify a loss, so your team gets the details to check. Some callers won't want to talk to a virtual receptionist at all. And its intake is only as good as the questions it's built around for your shop.

Whichever you pick, count again a month later the same way, and compare like with like: raw with raw, filtered with filtered, new callers with new callers. To hear what a fully answered call sounds like start to finish, press play on a real call on willisonhq.com. It's our receptionist answering for a sample HVAC shop. A restoration build runs the same engine around your own intake questions.

Frequently asked questions

It's close to the home services average, not a target. Invoca's 2026 report puts restoration at 51% across its restoration customers, a point below the 52% it found for home services overall. Whether that's good enough for your shop depends on what sits in the unanswered half. Misdials and supplier calls cost you little; a homeowner with water spreading across the floor may cost you the job.

No. Invoca publishes averages across its customer base and doesn't break them out by time of day, so there's no published night rate for restoration. To see your own, split a month of your call records into business hours and after hours and divide each slice separately.

Yes. If your after-hours line forwards to an on-call phone, those calls were answered by a person, and leaving them out makes your night rate look worse than it really is. Pull that phone's history for the same dates, keep only the calls that came through the forward, and match them to the main log so nothing counts twice before you do the division.

Not in Invoca's benchmark, which counts calls a person answered. On your own count, give AI-answered calls a column of their own. They aren't unanswered if the caller got through and the details reached your team, but they aren't the same thing Invoca measured either, and blending them in would make any comparison to 51% meaningless.