Per-call vs per-minute answering service pricing: which one costs you less
Two answering services quote you. One charges a flat fee every time they answer. The other charges for every minute they’re on the line. The numbers are in different units, so there’s no way to eyeball which is cheaper.
There is a way to settle it, and it’s one division. This post is that calculation, plus the fine print that can change the answer by more than the gap between the two headline rates. If what you want instead is the wider picture of what answering services cost and what drives the bill, that’s a separate guide.
The short version: divide the per-call price by the per-minute rate and you get a call length in minutes. That’s your break-even. If your calls usually run shorter than it, per-call wins; longer, and per-minute wins. Then check the two things that override the maths — the rounding increment and what counts as a billable call. Both matter more than they sound: across six providers checked for this post — five of them not us — published increments ranged from one second to thirty, and the same 65-second call bills as anything from 65 to 120 seconds depending only on which one you’re on. Half didn’t publish the rule at all.
What each model actually bills
Per-call charges a flat amount each time a call is answered, no matter how long it runs. A twenty-second message and a six-minute booking cost the same. Smith.ai bills this way, describing it as “billing per client call” with “no mystery accounting involved.”
Per-minute charges for connected time. Short calls are cheap, long ones aren’t, and the meter is the thing to understand: ask exactly when it starts and stops, because “connected” and “talking to a person” are not always the same moment. AnswerFirst, Moneypenny, PATLive and Ruby all price in minutes.
That difference is the whole decision. Per-call puts the risk of long calls on the provider; per-minute puts it on you. Which is better depends entirely on how long your calls actually run — which is a fact about your business, not about the provider.
The one calculation that decides it
Take the per-call price and divide it by the per-minute rate.
break-even call length = per-call price ÷ per-minute rate
The result is a number of minutes. That’s the call length where the two quotes cost exactly the same.
- If your average call is shorter than the break-even, per-call is cheaper.
- If your average call is longer, per-minute is cheaper.
So the question stops being “which model is better” and becomes “how long are my calls?” — which you can find out. If you’re already with a service, the answer is on your invoice. If you’re not, a fortnight of noting how long your own calls run is enough to get close.
Two things worth knowing before you trust the number. Averages hide their own shape — if most of your calls are thirty seconds but a handful run ten minutes, the average sits in the middle and describes none of them. Look at the spread, not just the mean. And the number moves with the job you’re asking the service to do: taking a message is short, booking an appointment is longer, qualifying a lead is longer again. If you’re about to ask for more than message-taking, calculate against the calls you’re about to have.
What providers actually do — six, checked
Rather than describe the market in general terms, here is what five answering services published on their own pricing pages on 25 September 2026 — plus our own, held to the same test:
| Provider | Billed by | Rounding rule on the pricing page? | Billable-call rule on the pricing page? |
|---|---|---|---|
| Smith.ai | the call | Not stated | Yes |
| AnswerFirst | the minute | Yes — one-second | Not stated |
| Moneypenny | the minute | Yes — 30-second | Not stated |
| PATLive | the minute | Not stated | Not stated |
| Ruby | the minute | Not stated | Not stated |
| Rindee | the minute | Yes — by talk time on plan minutes. The pay-as-you-go platform fee is per started minute; full rule here | Not stated |
Three things follow from that, and they’re the reason the rest of this post is a list of questions.
The increment genuinely varies, and it’s not a small difference. AnswerFirst advertises “True one-second billing (never rounded up)” and makes the contrast itself: a 65-second call, they say, “may cost 90 or even 120 seconds” elsewhere, where “it costs exactly 65 seconds here.” Moneypenny publishes a worked example of its own 30-second increments — “A 1-minute, 5-second call is billed as 1 minute and 30 seconds” — alongside what full-minute billing does to the same call: “calls are rounded up to the next whole minute… A 1-minute, 5-second call is billed as 2 minutes.”
So the same 65-second call bills as 65 seconds, 90 seconds, or 120 seconds depending only on the increment. That’s up to 84% more for an identical call, before anyone’s rate enters the calculation.
The pricing page often isn’t where the rule lives. Half of these six state no rounding rule there, and five of the six — ours included — say nothing on that page about what counts as a billable call. That isn’t an accusation, ours least of all; these rules frequently sit in terms or an FAQ instead, or nowhere public at all. Which is the point: this is something to ask for rather than expect to read.
Where the billable-call rule is published, it has conditions. Smith.ai, the per-call provider here, states plainly that “you won’t be charged for spam or unsolicited sales calls” — and attaches a condition: “as long as the lines you forward to us pass Caller ID.” Even the clear answer has a clause in it.
The fine print to pin down
The break-even assumes both quotes mean what they say. Each of these is a question to ask, in writing, before comparing anything:
- How is time rounded? The big one on per-minute pricing, and the one least likely to be on the page — three of the five above don’t state it. Ask for the increment in seconds, and ask whether it rounds up or to the nearest. The shorter your calls, the more it matters: rounding is a fixed penalty per call, so it hurts a thirty-second call far more than a five-minute one.
- Is there a minimum billable length? A minimum per call does the same thing as rounding, by a different name. None of the five published one either way.
- What counts as a billable call? On per-call pricing this decides everything. Wrong numbers, hang-ups after two seconds, telemarketers, and repeat calls from the same customer within a few minutes — find out which of those you pay for. Per-call pricing puts no billing incentive behind filtering them out, so get the rule in writing rather than assuming it works in your favour.
- When does the meter start and stop? Ask specifically about hold time, menu time, and the wrap-up after the caller has gone. Time billed while nobody is speaking to your customer is time you’re paying for nothing.
- Is there a monthly minimum or commitment? A floor you pay in a quiet month can swamp the difference between the two models entirely.
- Do the rates change outside business hours? If a surcharge applies overnight and at weekends, run the break-even separately for those hours — the answer can flip.
Work through those six before you compare, because any one of them can move the real number further than the difference between the two headline rates.
Which model suits which business
The break-even does the arithmetic, but the pattern behind it is simple enough to sanity-check against:
Per-call tends to suit businesses whose calls are short and uniform — a service that takes a name, a number and a reason, and passes it on. Predictable per call, and long calls aren’t your problem.
Per-minute tends to suit businesses with low or erratic volume, where paying for exactly what you use beats paying a flat fee per answer — and businesses whose calls are genuinely long, where a per-call price high enough to cover a ten-minute booking would be punishing on the short ones.
The awkward case is a business with both: lots of quick calls and a few long ones. There, the average lies to you, and the honest move is to price both quotes against your actual call-length distribution rather than a single number.
The question that sits underneath both
Both models bill for answering. Neither bills for outcome.
Worth holding onto when you compare, because a message taken still leaves the job undone — someone on your side has to call back, and the customer waits in between. A call that ends with the appointment in your calendar is worth more than one that ends with a note about it, at any per-minute rate. When you’re comparing quotes, compare what the call finishes, not just what it costs.
How to compare two quotes properly
- Get both in writing, including increment, minimums, and what counts as billable.
- Work out your average call length and, if you can, the spread.
- Calculate the break-even — per-call price ÷ per-minute rate.
- Compare it against your average, and against your typical short call and long call separately.
- Normalise both to a monthly figure at your real volume, with surcharges and minimums added in. That’s the only number that’s comparable.
- Ask what each quote does when the caller wants an appointment, not a message.
Billed by talk time, and pay-as-you-go by default. Rindee deducts plan minutes by the talk time actually used rather than in rounded-up blocks — though the pay-as-you-go platform fee is charged per started minute, so apply this post’s own advice to us and read the full rule. Pay as you go is the default — a small platform fee plus usage from a prepaid balance, no plan required and no contract. Optional plans exist if you’d rather a predictable monthly bill, and if you use up an included bucket the extra bills at the pay-as-you-go rate so calls never stop. Inbound is never capped. Current rates are on the pricing page, and calls end with the appointment booked rather than a message for someone to action later.