Is an answering service worth it for my small business?

It depends on one number, and it is not the rate. An answering service is worth it when the work it recovers is worth more than the minutes it burns. BotsDontSleep charges $0.60 CAD per answered minute, billed by the second, with $0 setup and no monthly minimum, 24/7. A small business taking 100 calls a month that average two minutes pays about $120. If one job is worth $200 to you, that bill is covered by recovering a single job every second month. If you already answer nearly every call yourself and nothing rings out at night, the honest answer is that it will not find you money.

What does an answering service cost a small business a month?

Minutes multiplied by a rate, and with BotsDontSleep nothing else. $0.60 CAD per answered minute, billed by the second, $0 setup, no monthly minimum, no per-seat fee, no contract. A hundred calls averaging two minutes is 200 answered minutes, or about $120.

Live, human-staffed services price the same way and add fees around it. Published talk-time rates commonly run $0.75–$1.50 USD a minute, per the Alliance Virtual Offices pricing guide published 21 July 2026, and cost guides put typical small-business monthly plans between $135 and $400, per Housecall Pro's guide published 11 May 2026 and updated 30 July 2026. Answering365's guide, published 8 June 2026, lists one-time activation fees of $50–$500, patch or transfer fees of $0.50–$2.00 per call, and script-change fees of $25–$100. Those are published figures on those dates in the currency each guide quotes; check each provider's own site for current numbers before you compare anything.

The full rate comparison lives on our answering service cost page. What a rate card cannot tell you is whether the number is worth paying, which is a different question and the one you actually have.

How do I work out whether it pays for itself?

Four steps, about twenty minutes, and you already have most of the inputs.

1. Pull last month's call log from your phone provider. You want inbound call count and, if the log carries it, duration. Not a number from memory — owners consistently underestimate how many calls ring out while they are working.
2. Work out your monthly cost. Answered calls multiplied by average call length, multiplied by $0.60. If your log has no durations, use two minutes as a planning figure and correct it after your first month of real transcripts. We are treating two minutes as an assumption here, not a measurement.
3. Decide what one job is worth. Your average invoice is the conservative version. If customers come back, first-year value is the truer one, and it is usually two or three times larger.
4. Divide the cost by the job value. That gives you the number of recovered jobs a month that puts you level. Anything above it is profit; anything below it means the service is a convenience purchase, not a return.

The reason this arithmetic is worth doing on paper is that the answer is rarely close. It either clears by a wide margin or it does not clear at all, and which one you are looking at is usually obvious by step four.

What does break-even look like at my call volume?

Using $0.60 CAD per answered minute, billed by the second. The two right-hand columns show how often you would need to recover one job that would otherwise have gone elsewhere.

Two honest notes on that table. The job values are yours to supply — $200 and $500 are placeholders, and a roofing quote and a haircut sit nowhere near each other. And the table shows a break-even, not a forecast. It tells you how small the recovery has to be before the maths works. It does not promise the recovery.

What it does show is where the pressure sits. At 25 or 50 calls a month the bill is small enough that almost any recovery clears it, which is why low-volume trades tend to find this obvious. At 400 calls a month the bill is real money and the question becomes how many of those calls were being handled badly in the first place.

What is a missed call actually worth to me?

Nobody can tell you that from the outside, and the industry averages you will find quoted are worth very little when applied to one business. Here is how to get your own figure instead.

Take your call log for a full month and count three things: calls that rang out with no answer, calls that went to voicemail with no message left, and calls that were answered after four rings or longer. Then look at how many of those numbers called back. The ones that never called back are your real number — not lost calls, but lost callers.

Our founder's own version of this arithmetic is on the about page: a friend's home-services company was losing roughly $200,000 a year to a phone nobody could get to. That is one business with high job values and heavy after-hours demand, and it is a reason to run your own numbers rather than borrow his.

The two figures that move this most are job value and how urgent your callers are. A caller with a burst pipe at 11 p.m. dials the next name on the list within a minute of hearing voicemail. A caller comparing quotes on a kitchen renovation might wait until Tuesday. Same missed call, entirely different cost.

Is it worth it if I already answer most of my calls?

Sometimes, and for a reason that has nothing to do with missed calls.

If you answer 95% of your calls, the recovery argument is weak and you should be honest with yourself about that. But answering the phone yourself has a cost that never appears on an invoice: you take the call while you are under a vehicle, in someone's basement, or mid-consultation, and the interruption costs you more than the call is worth. A shop owner taking twenty calls a day is being pulled off paid work twenty times.

That is a different purchase. It is worth it when the hour you get back is billable and the calls you are fielding are mostly sorting — where are you, are you open, is my car ready. It is not worth it when the calls are the relationship, which is the next section.

There is also the volume problem that shows up in bursts rather than averages. A storm, a marketing push, or one busy Monday produces four calls at once, and a person can answer one of them. Concurrency is unlimited here and unmetered, so the fourth caller is greeted at the same speed as the first. If your call volume is flat and never spikes, that capability is worth nothing to you and you should not pay attention to it.

When is an answering service not worth it?

Five cases, stated plainly, because the arithmetic above is easy to make say anything.

You take fewer than about twenty calls a month and answer nearly all of them. The phone in your pocket is already doing the job. At that volume the service costs almost nothing, which also means it recovers almost nothing.

Your enquiries arrive by chat, form, or email. BotsDontSleep is a voice product. It answers the phone and it does not run your web chat, so if your leads are not phoning, this fixes a problem you do not have.

Your customers are paying for a relationship with one named person. A consultant, a therapist, a solo practitioner with a book of long-standing clients — being answered by anything other than that person is a downgrade, however well it is done. Our long-tail industries page is candid about which one-person businesses this suits and which it does not.

Your calls are genuinely long. Per-minute pricing favours short calls. If your average intake runs fifteen or twenty minutes, a flat monthly plan may cost less for the same work, and you should do that arithmetic before switching. We break down where each pricing model wins in how much an AI answering service really costs.

You cannot service the work you are already winning. This is the one people skip. If you are three weeks behind and turning jobs down, an answering service books more work you cannot do and converts a phone problem into a scheduling problem and an angry-customer problem. Fix capacity first.

One more limitation that applies even when it does pay: an answering service recovers calls, not customers. It cannot make someone choose you, it cannot rescue a bad quote, and it will not repair a reputation problem. It removes one specific failure — the caller who could not reach anyone — and leaves every other reason you lose work exactly where it was.

How do I test whether it's worth it without committing?

Forward one line and count.

Point your after-hours number at the service and leave your daytime line alone. Forwarding takes about ten minutes at your carrier, and unforwarding takes the same ten minutes. Setup is $0 and there is no contract or minimum, so a month of testing costs the price of the minutes it answers, which for most after-hours-only pilots lands under $50.

Then read the transcripts. Every call is recorded, transcribed, summarized, and scored, so at the end of the month you are not estimating anything — you have a list of calls that were answered, what each caller wanted, and which ones turned into work. Set the bookings against the bill. That is the only version of this calculation that uses your numbers instead of somebody's averages.

Most businesses are live in about one week: roughly a day for discovery and number connection, three to five days for tuning and test calls you read yourself, and same-day launch once you approve. If you would rather compare it against a part-time hire before any of that, the arithmetic on wages and coverage hours is in AI receptionist vs human receptionist.

Or dial it and decide in ninety seconds. The demo line is +1 437-494-9110. Try booking something, then try asking it a question it cannot possibly know, and see whether it admits that or invents an answer.