Flat rate or per minute: which answering service pricing model is cheaper?

Your answering service bill made sense in April. Then the first real heat wave hit, the phone went off for nine days straight, and the invoice came in at a number you had to read twice. Maybe the meter found a busy month. Maybe you ran through an included block and out the far side. Either way, nothing about the service changed. That's the part you don't think to ask about when you're being quoted a rate.

Here's the short version. Neither model is cheaper on its own. Per minute is usually cheaper when your call volume is low, your calls are short, or you have a long quiet season. Flat rate is usually cheaper when your volume stays high enough to use most of what you pay for, month after month, and it's the more predictable bill when volume swings, which in the trades it tends to. And the comparison that answers the question is the twelve-month total, not one month's invoice, because a model that wins in July can still lose across the year.

One thing to get out of the way first, because it should change how you read the rest. Willison bills flat: one monthly price, scoped up front to your call volume and needs, with no per-minute meter, no overage, and no surcharges, so a busy month or a whole busy season doesn't move the bill. It isn't infinite, though, and we'd be doing exactly what this post warns about if we said it was. If a business grows into a genuinely bigger operation, we scope the price again, with 30 days notice and the option to cancel first.

So we have a side in this. What follows is the argument we'd make to another owner, including the parts where per minute is the better buy and where flat rate stings.

Which of the three are you actually being quoted?

Per minute, you're buying time: somebody picks up, a clock runs, you pay for what it records. Flat rate, you're buying capacity: a set amount for a defined block, inside which one more call costs you nothing.

A hybrid is both. A monthly base that includes some number of minutes or calls, then a rate for everything past it, which is a block of capacity with a meter bolted to the far end. If that's what you're holding, the decision isn't flat against per minute at all. It's which tier you sit in, how much headroom it gives you, and what the meter charges once you're through it.

What runs up a metered bill that you won't think to ask about?

Five things, and only one of them is talking to customers. None of this is universal, which is exactly why each one is a question rather than a warning.

Ask what the billing increment is. A rate quoted per minute isn't necessarily measured by the second. If time rounds up to the next full minute, a thirty second wrong number costs the same as a full minute of real work. Ask this before you ask the rate, because the increment can matter more than the rate does.

Ask where the monthly minimum sits. A minimum is a floor, not a discount. In a quiet month you pay it and use a fraction of it, so on a low-volume line the minimum, not the rate, is effectively your bill.

Ask what counts as a billable call. A meter can't tell a homeowner with a dead furnace from an auto-dialer. Find out whether robocalls, wrong numbers, and four-second hangups are excluded, and what happens to a call that connects and dies immediately.

Ask whether outbound minutes and wrap-up time bill. This is the one that surprises trades owners, and on an after-hours rotation it can be the biggest line on the invoice. One 2am call isn't one call. The service answers the homeowner, rings your on-call tech, gets no answer, rings the second tech, then calls the homeowner back to confirm. That's one job and four separate stretches of metered time, plus whatever it takes to type up and send the message afterward.

Ask what a typical call actually runs. Not your average handle time, theirs. A scripted intake plus wrap-up can run longer than the way you answer your own phone, and their average billed minutes per call is the number your invoice is really made of. A service that won't give you that number straight has told you something.

We took the screening side of this apart separately, in whether an AI receptionist can filter spam without dropping real customers.

Where does flat rate burn you instead?

In four places, and a flat-rate pitch won't volunteer any of them.

The overage rate. This is the big one, and it's the exact thing a flat plan is sold as protecting you from. On a plan with an included block, you're back on a meter once you're past it. Ask for that rate as a number, and ask how it compares to the blended rate you're paying inside the block, because it can sit well above it. A plan that absorbs a busy month up to a line and then prices the far side punitively hasn't absorbed anything, it's just moved where the cliff is.

Paying for July all year. Capacity is sized for your peak. You then carry that same monthly number through October, November, February, and March, when the phone is quiet and you're using a fraction of what you bought. For a shop with a genuine off season, that's the whole argument against flat rate, and it's a real one.

What happens to what you don't use. Ask whether unused minutes or calls roll over, and if they do, how long they live. If they don't, every quiet month is capacity you paid for and threw away.

Getting stuck in the wrong tier. Ask what it takes to move down a tier as well as up, and whether you can do it mid-term. A plan sized for the season you just had is the wrong plan for the one coming if nothing lets you change it.

Put those beside the metered list and the shape of the decision gets clearer. Per minute bills you for the month you actually had. Flat rate bills you for the month you planned for, and on a plan with an included block, meters you anyway once you're past it.

Why do the trades make this harder than it looks?

Because trades volume doesn't arrive evenly, and both models quietly assume a steadier phone than you have.

An HVAC phone is calm in October and unrecognizable the first week it hits the high nineties. A roofing phone follows the weather, and the week after a hail line comes through is the week every call in the county is up for grabs. Restoration is the sharpest version of it, because water losses cluster on the cold snaps and the storms rather than spreading themselves evenly across the year.

Here's the part that catches owners out. Those are the same weeks your revenue spikes, so a bigger bill feels fair. But a metered bill scales with ring volume, not with booked jobs, and a peak week's ring volume is thick with price shoppers and people working down a list.

You get billed for the whole storm. You only get paid for the part of it you booked.

Why your own talk time will understate the bill

This is the trap in estimating a metered quote from last year's minutes. Your historical talk time only counts the calls somebody actually picked up. A service answers the ones you're currently missing too, and those become billable minutes that have never appeared on any record you own.

How big that gap is depends on your trade, and it varies more than most owners expect. Across home services, Invoca's 2026 benchmarks put the answer rate at 52% of calls.

By sub-industry it splits hard. The same report puts HVAC at 34%, the second lowest of the nine measured, against 74% for plumbing.

Run that forward. An HVAC shop answering about a third of its calls today, moving to a service that answers all of them, could be looking at up to three times the answered calls its own logs show. A plumbing shop already answering three quarters is looking at closer to a third more. Same rate card, very different invoice, and the shop with the worse answer rate gets the bigger jump. The same jump applies to a plan with an included block, so size any tier against the full count too, not against what your logs show you answered.

Invoca also reports the answer rate rising to 65% for calls over 15 seconds and 73% over 30 seconds, thresholds it says filter out misdials and quick hangups. So some of what you're missing is noise rather than customers. That doesn't shrink the estimate, though, because a service answers the misdials too, and unless they're excluded, a meter bills for them.

So how do you actually compare two quotes?

Two passes. The first answers which is cheaper. The second answers whether you can live with it.

Pass one: build the twelve-month total. Pull your monthly inbound call counts for the last year off your carrier log or call tracking. That count already includes the ones nobody picked up, so use it as it stands and don't scale it up. If all you have is talk time, don't build the estimate from it: take last year's answered-call count, divide by your answer rate to estimate total inbound calls, and then use the vendor's average billed minutes per call rather than your own minutes.

One adjustment runs the other way, whichever count you start from. A caller who didn't get through often dials again, and a service that picks up the first time never sees that second call, so count unique callers rather than total rings where your log lets you.

Then ask the vendor for its average billed minutes per call. Ask whether that figure is already rounded to the billing increment, so you don't apply the increment twice. If it isn't, ask for the rounded version, because rounding each call up is not the same as rounding the average.

Ask too whether it includes outbound legs and wrap-up. If it doesn't, get those figures from the vendor and add them in yourself, because on an after-hours rotation they're the part that moves the total. For a metered quote, multiply calls by billed minutes by the rate, and in any month where that comes in under the monthly minimum, use the minimum instead. For a plan with an included block, start from its monthly price and add the overage rate times whatever runs past the block.

Do that for both quotes, month by month, and total all twelve. Whichever total is lower is the cheaper plan, and the rate line on its own won't tell you which.

A metered plan that stings in July can still win the year on a quiet autumn, and that's a legitimate way to win.

Pass two: look at your worst month on its own. Not to decide which is cheaper, but to find out what the ugliest invoice of the year looks like and whether it lands in a month you can absorb. That's a cash-flow question, and it's a fair reason to pay more across the year for a number that doesn't move.

One thing neither pass measures: what the service actually does on the call. Taking a message and booking the job aren't the same product, so decide which you're buying before you compare prices at all. The number that makes two genuinely different services comparable is what you pay per booked job, not what you pay per minute.

When is per minute genuinely the cheaper choice?

More often than a flat-rate pitch will admit, and we'd rather say so than pretend otherwise.

If your phone rings a handful of times a week after hours and you mainly want a name and a number taken, per minute is honest pricing. You use little, you pay little, and you aren't funding capacity you'll never touch. Check the minimum before you commit, though, because at genuinely low volume the minimum is your real price, so compare it against the flat quote directly rather than comparing rates.

Same goes for a true backup line, the one that only rings when your front desk is at lunch or already on another call. And for a shop with a real off season, paying almost nothing through the quiet months can be worth a rougher peak. We looked at that trade-off for a smaller operation in whether round-the-clock answering is worth it for a small plumbing business.

Per minute isn't a trap. It's a bet that your minutes across the whole year stay low, and plenty of shops make it and win. Run your own twelve months before you assume you're not one of them.

The line item that decides the rest

Whichever model you sign, the bill is only half the question. The other half is how many of those calls turned into work, and that half never appears on an invoice.

Willison answers every inbound call, 24/7, in seconds, effectively on the first ring. It answers many of them at once, so the tenth caller on the first hot morning of the year gets picked up as fast as the first one did, with nobody on hold and nobody hearing a busy signal. It asks your intake questions, qualifies the job, and books it straight onto the calendar.

Being straight about the limits. Answering everything also means answering more robocalls and more wrong numbers, and nothing on the phone builds you another truck. If your calendar is already full by nine on a July morning, faster pickup buys you better information about what's waiting, not more capacity. Some callers just want a person too, which is what an escalation contact is for.

If you want to hear what one of those calls actually sounds like, press play on a real call on willisonhq.com and follow it from first ring to booked visit.

Frequently asked questions

It's the smallest unit of time a service bills in, and it matters because it sets the floor under every single call. Billing can be measured by the second, or rounded up to some fixed block, or rounded up to a full minute, and the practice varies from one service to the next. On a line that takes a lot of short calls, that choice can move your invoice more than the headline rate does, which is why two services quoting the same rate can bill very differently. Get it in writing before you compare quotes.

Assume yes until somebody tells you otherwise in writing. A meter doesn't know the difference between a homeowner with no heat and an auto-dialer, so the safe assumption is that a call somebody picked up is a call you'll be billed for. Ask directly whether spam, wrong numbers, and quick hangups are excluded, and ask what happens to a call that gets answered and disconnects four seconds later. On a plan with an included block, junk calls don't show up as a separate charge, but they still eat the block and push you toward the overage rate sooner.

No, and the two get confused constantly. Per call is still usage billing: you pay for each call that comes in, so a busy month still costs you more than a quiet one. A flat rate is a fixed amount for a defined block, and inside that block one more call costs nothing extra. Per call looks like a simpler version of flat rate and behaves like the metered one, so run it across twelve months the same way you would run per minute.

Start with the rate, then get the things that decide what that rate actually costs you. What billing increment is it measured in. What's the monthly minimum, and does anything unused roll over. What counts as a billable call, including robocalls and hangups. Whether outbound legs and wrap-up time bill. How big the included block is and what the overage rate is past it. What the service's own average billed minutes per call runs. What the term is and what it takes to cancel. And what it does on the call, take a message or book the job. The rate on its own tells you almost nothing until you have the rest beside it.