How much is your business losing to missed calls and no-shows?

The slider is only as good as the number you put into it, and most owners guess low. Here is how to get a real figure rather than an impression.

1. Open your phone’s call log and filter to the last 30 days. On most Android handsets, missed and rejected calls are already a separate tab. On a business landline, ask your telecom provider for a call detail record — they are obliged to give it to you.
2. Count missed calls, not total calls. A missed call is one that rang out, hit a busy tone, or reached voicemail. Exclude your own outgoing calls and anyone in your contacts, because those are usually staff and suppliers rather than customers.
3. Split them by hour. Mark which ones arrived while you were open and which arrived outside your staffed hours. This split matters more than the total, because after-hours calls are the ones nobody is ever going to call back.
4. Remove the obvious repeats. The same number ringing four times in six minutes is one frustrated person, not four enquiries. Count it once.
5. Ring twenty of them back. This is the step people skip and the only one that produces a real conversion rate. Ask what they wanted and whether they got it elsewhere. You will get a number specific to your business rather than our 35% assumption.

That last step usually changes the estimate more than any slider. If eighteen of your twenty callers say they went somewhere else, your loss is far above the model. If most say they were price-shopping and never intended to book, it is far below.

Three numbers do the work in this model, and none of them are research findings. They are our working assumptions, shown openly so you can replace them with your own. A calculator that hides its assumptions is a lead-capture form wearing a maths costume.

The 35% conversion assumption. We assume roughly a third of people who could not reach you go on to book with someone else. This is too high for a business whose customers have no substitute nearby, and too low for anything where the next option is one search result away — a restaurant, a salon, an emergency plumber. If you did the twenty callbacks above, use your own figure.

The 18% no-show assumption. No-show rates vary enormously by sector and by whether you take a deposit. A clinic that takes ₹200 up front and sends a reminder the night before runs far below this. A free-consultation business with no reminder runs above it. If you already track no-shows, use your real rate.

The ₹120/hour staff cost. This is a rough blended cost for front-desk time at an Indian small business, and it will be wrong for your city and your salary structure. Take the monthly cost of the person who answers your phone, divide by the hours they work, and use that instead.

The recovery side has assumptions too, and they are optimistic on purpose so you can discount them: the model credits an AI agent with recovering all missed bookings, all the staff time spent on repetitive questions, and 80% of no-show revenue. Nothing recovers 100% of anything. If you halve every recovery figure and the number is still worth acting on, it is worth acting on.

These are illustrations of the model at typical inputs, not measurements of real businesses. Run your own numbers in the calculator above — the point of the table is to show how much the ticket size matters compared with the call volume.

Each row is calls/day × 30 × miss rate × ticket × 35%. The dealership and the clinic dominate not because they miss more calls but because each lost call is worth several times more. If your ticket size is small, the honest conclusion is that missed calls cost you less than this page’s headline suggests, and you should weigh the fix accordingly.

A single monthly rupee figure hides the fact that missed calls have four quite different causes, and a fix that solves one does nothing for the others. Before you buy anything, work out which of these is actually yours — the hour-by-hour split from the counting exercise above usually makes it obvious.

1. Nobody was there. The call arrived outside your staffed hours — evenings, Sundays, festivals, the two hours the desk is unmanned at lunch. This is usually the largest bucket and the most under-counted, because these callers rarely try twice and never leave voicemail. Forwarding to a second handset does nothing here, since the second handset is also off. What fixes it is something that answers when nobody is there at all: an after-hours answering setup, or accepting the loss and pricing it in.

2. Everyone was busy. The call arrived during your peak, when the one person at the desk was already on another line or with a customer in front of them. This bucket clusters tightly — you will see it in the log as three or four missed calls inside twenty minutes, usually at the same time each day. The cheap fix is a second handset or a rota change at that specific hour. The structural fix is something that can hold more than one conversation at a time, since a queue only drains as fast as a human frees up.

3. The caller gave up on the menu. If you run a phone menu, some share of callers hang up during it rather than after it. These are invisible in a missed-call count because the call was technically answered — by the menu. If your log shows very few missed calls but your enquiry volume feels low, this is the likely explanation, and the honest comparison of menu versus conversation is in AI voice agent vs IVR.

4. The caller could not be understood, or could not understand you. A caller comfortable in Telugu, Tamil, Marathi or Bengali who reaches a desk operating in English will often end the call quickly and politely, and it will look in your log like a short answered call rather than a lost one. This bucket is invisible to every metric on this page. The only way to find it is to listen to a few short calls or ask staff how often it happens.

Buckets one and two are what an always-on answering setup is genuinely good at. Bucket three is a configuration decision you can make today. Bucket four is a language question rather than a phone question. Spending money on the wrong bucket is the most common way this calculator leads people astray.

Work down this list in order. The first three cost nothing and fix a surprising share of the leak, and we would rather you tried them before paying anyone — us included.

1. Put your actual hours on your Google Business Profile. A large share of “are you open?” calls disappear when the answer is already on the listing that brought them to you.
2. Put your prices, or a price range, on your website. Price-shopping calls are the easiest to eliminate and the least valuable to answer.
3. Turn on call forwarding to a second handset. If two people can pick up instead of one, the busy-tone share drops immediately at zero cost.
4. Send a WhatsApp reminder the day before every appointment. This addresses the no-show line of the calculator rather than the missed-call line, and it is often the larger number.
5. Take a small deposit on high-value bookings. Even ₹100 changes no-show behaviour more than any reminder.
6. Only then consider an AI receptionist. It earns its keep on the calls the steps above cannot reach: simultaneous callers, after-hours enquiries, and callers speaking a language your staff does not.

If steps one to five close the gap, you have your answer and it cost you nothing. We would rather tell you that than sell you an agent you did not need — the detail on where an AI agent genuinely does and does not pay for itself is in is an AI receptionist worth it.

The calculator will always produce a positive number, because multiplying four positive sliders cannot do anything else. That is its main weakness, so here is the counterweight.

If you miss fewer than about three or four calls a week and someone is nearly always at the desk, you do not have a phone problem — you have a normal business, and the money is leaking somewhere else. If your average ticket is small, recovering a handful of calls a month will not cover the cost of any paid fix. And if your calls are overwhelmingly existing customers rather than new enquiries, the “they booked with a competitor” assumption does not apply to most of them, so the real loss is a fraction of what the model shows.

A good test: halve every recovery assumption, then ask whether the remaining number is larger than the annual cost of the fix. If it is not, close the tab. Nothing on this page is worth buying on the strength of a slider.

Whatever you change, measure the same four things you measured before you changed it, over the same length of window. Most people never do this and end up arguing from impressions.

• Missed calls per week — from the same call log, filtered the same way. This should fall first and fastest.
• Answered-within-30-seconds share — speed matters more than total answer rate, because a caller who waits does not stay. The reasoning is in speed to lead.
• Bookings made outside staffed hours — if this is still zero after a month, whatever you bought is not covering evenings and Sundays.
• No-show rate — track it separately from missed calls. Reminders and deposits move it; answering the phone does not.

Give any change a full month before judging it, and compare like-for-like weeks — a festival week or an exam week will distort a fortnight’s data badly. If you want the definition and the wider context of the metric itself, see missed call alert.