How do I find out how many calls I actually miss?

Your mobile keeps a call log, and most business phone plans or VoIP services will give you a call report on request. Export a fortnight of it, then sit down with the list and mark each missed number: supplier, existing customer, someone who called back and got through, or a new enquiry you never spoke to. The last category is the only one that matters, and it is usually smaller than the raw missed-call count and more expensive than the guess.

Is an automated text back to a missed call going to annoy people?

Not if it is honest and short. What annoys people is a message pretending to be you when you are clearly on a roof, or one that asks nothing and promises nothing. A message that says you are on a job, gives a time you will call back, and asks what the work is, reads as a business that is running properly. Send it from the number they rang, and make sure the callback actually happens.

Should I just hire someone to answer the phone instead?

Run the number first. If the missed enquiries are worth more per year than a part-time person costs, a person is the better answer, because they can qualify the job, book it in and handle the awkward ones. If the number is smaller than that, you want the cheapest thing that stops the caller ringing the next business, not a hire you will resent in six months.

What a missed call actually costs a trade business

Why does a missed call cost more than a call back?

Because of when it happens. A person rings a trade at the moment the problem becomes real: the hot water has gone, the tenant has complained, the job needs to start before the slab pour. They are not researching. They are buying.

At that moment the caller has a short list, usually pulled off a search result or a mate's recommendation, and they are ringing down it. The first business that answers with something credible gets to have the conversation. Everyone else gets to leave a voicemail nobody listens to.

The reason it hits trades harder than most is structural. The person best placed to answer is on the tools, up a ladder, under a house or driving. The busier the week, the more calls get missed, which means the leak is worst exactly when the business looks healthiest. That is why it rarely gets noticed: it does not show up as a bad month, it shows up as a month that could have been better.

How do you put a number on it?

Four numbers. Get each one from what you already have rather than from a guess.

Missed new enquiries per week. Not total missed calls. Export a fortnight of your call log, mark every missed number as supplier, existing customer, repeat attempt, or genuine new enquiry you never spoke to, and count only the last group. Halve the fortnight to get a week.
The rate you would win them. Use your own quote-to-job conversion, which you can pull from your accounting or job-management file: jobs won divided by quotes sent over the last few months. Then knock it down, because a caller who reached you second is less likely to buy than one who called you first.
Average job value. Total invoiced over a recent period divided by the number of jobs. Use the median if one big job would distort it.
Weeks worked per year. Your actual figure after holidays and the quiet stretch.

Multiply the four together. That is the annual revenue walking to a competitor before you have spent a cent trying to fix it.

Do the arithmetic on paper with your own numbers. Even at modest values the total tends to land somewhere between "that is a decent ute" and "that is a second van and the person to drive it", which is why the exercise is worth the fortnight it takes.

Which missed calls should you not count?

Most of them, honestly. Overstating this number is the fastest way to talk yourself into buying something you do not need.

Do not count suppliers, the accountant, a recruiter, or a scam call. Do not count an existing customer who rang twice and got you on the third go. Do not count the same new number three times in ten minutes, because that is one enquiry, not three, and the fact they kept ringing tells you they wanted you specifically.

What you are left with is the honest figure: people who wanted to give you work, could not reach you, and never came back. It is a smaller list than the phone shows and it is the only one worth acting on.

What does the arithmetic leave out?

Four things, all of which push the real cost up rather than down.

Repeat work. A first job for a good customer is often the start of a decade of small ones. Losing the first call loses the tail, and the tail is where the margin usually is.

Referrals and reviews. Customers you never had cannot recommend you or leave you a review, and in most trades those two channels do more work than any advertising.

The evening tax. Calls returned after six are calls returned instead of dinner. That cost never appears in the accounts and is the one that eventually makes people sell the business or hate it.

The quoting queue. A call answered late becomes a quote written late, which becomes a quote never chased. The leak compounds through the rest of the process, which is the point of the wider list in five workflows trade and retail businesses leak time on.

What actually fixes it?

Take the cheapest option that stops the caller ringing the next business. Work down this order and stop when the number is covered.

Change who answers. Divert to whoever is closest to a desk, or split the number so quotes go one way and existing jobs another. Free, and it fixes a surprising share of the problem in businesses with any office presence at all.
A missed-call text back. The caller gets an honest message within a minute: you are on a job, here is when you will ring back, what is the work. It costs almost nothing, it holds the enquiry for the twenty minutes that matter, and it gives you the job details before you call.
A human answering service. When the calls are urgent, emotional or need triage, a trained person beats anything automated, and it is worth paying for. Check current rates yourself and price it against the annual figure you calculated, not against how it feels.
An enquiry route that is not the phone. A short web form or a text line, with the reply drafted for you and sent once you approve it, catches the people who would rather not ring at all.
Only then, anything that answers calls on your behalf. It is the most complicated option and it is the one that damages the business if it gets a job wrong, so it earns its place last and only if the first four have not closed the gap.

Whatever you put in, keep a person on anything a customer sees before it goes. Draft and approve is slower to build and much harder to embarrass yourself with, and it is worth knowing how the thing behaves when it breaks before you rely on it.

What to do next

Do the fortnight. Export the call log, mark up the missed numbers into the four categories, and count only the genuine new enquiries you never spoke to. Then run the four-number arithmetic and write the annual figure on a piece of paper.

If the number is small, you have just saved yourself from buying a solution to a problem you do not have. If it is large, start at the top of the list above and change one thing, then count again a month later to see whether it moved. That before-and-after count is the whole point, and it is the same test Shift applies to any build: if the number did not move, it did not work. If you would rather see the whole map priced before choosing, that is what a Flow Map is for.