What missed calls are actually worth

Somebody has probably shown you a number. Missed calls times your average sale times a conversion rate somebody picked, annualised, printed in bold. It is usually large, it is usually presented as money you are losing right now, and it is almost always built the same four ways.

The arithmetic is not the problem. Multiplication is fine. The problem is what gets multiplied, and in particular what gets skipped.

The common shape asks for four things: missed calls, average sale, a conversion rate and sometimes a recovery rate. Then it multiplies. Nextiva publishes one beside its voice-agent offer, and there are dozens more; the category is crowded and the shape is nearly always this.

Four boxes force four assumptions, and none of them is stated:

1. Every missed call was a different person.
2. Every one of those people was going to buy something.
3. None of them would have come back to you on their own.
4. Revenue is the thing that pays for the fix.

All four are wrong in the same direction, which is upwards. Correcting them is not pessimism. It is the difference between a number you can take to whoever controls the money and a number that will not survive the first question they ask.

Your phone system counts rings. A business case needs people.

A customer who needs a plumber does not ring once and shrug. They ring, get nothing, ring again twenty minutes later, and try somebody else in the afternoon. That is one lost job and three missed calls. Feed the three into an opportunity count and you have tripled the answer before any other assumption has run.

300 missed calls / 1.5 attempts per person = 200 people

The repeat rate is worth measuring rather than assuming, and it is one of the few numbers in this whole exercise you can actually get: pull a month of missed calls with the numbers attached and count the duplicates. If you have never done it, the honest input is "I do not know", and a calculator that will not accept that answer is a calculator that will invent one.

The second correction is the one people accept immediately and then forget to apply.

In any month of missed calls there are wrong numbers, automated calls, a supplier chasing an invoice, a courier at the wrong address, and existing customers ringing about a job that is already finished and paid for. None of them were ever a sale. In a small service business the qualified share is often well under three quarters, and you can estimate it from a week of answered calls rather than guessing.

This is also where a phone system earns its keep in a way that has nothing to do with revenue: when everyone is with a customer, what a caller hears decides whether they are still a prospect ten minutes later.

Here is the correction that does the most damage to the headline figure, and the one almost nobody applies.

Some share of the people you miss ring again the next morning. Some of them leave a voicemail you already return. Some of them are booked in by the end of the week without anybody doing anything differently. That revenue is in your accounts today. It is not waiting to be recovered, because it was never lost.

A model that does not subtract it first is claiming credit for money you already have. And the subtraction has to come before the recovery rate, not after, because whatever you change can only reach the people who are still unrecovered:

still unrecovered = qualified opportunities x (1 - the share who return on their own) newly recovered = still unrecovered x the share your change would reach

Get that order wrong and you count the same caller twice: once as a customer who came back by themselves, and once as a customer your new setup rescued.

The last correction is the one that decides whether the case is real.

A recovered job brings in the sale value and then costs you materials, labour, fuel and time. What is left is contribution, and contribution is what has to pay for whatever produced the recovery, whether that is software, an answering service or somebody's extra hours. Comparing monthly revenue against a monthly cost compares a number that has not paid its own bills against one that has.

Once you are in contribution, a more useful question appears, and it is the one to put in front of a sceptical partner: how many extra jobs a month does this have to win before it pays for itself? If a job leaves you 190 dollars of contribution and the change costs 180 dollars a month, the answer is one. One extra job. That is a far easier thing to believe or disbelieve than a five-figure annual recovery estimate, and it is the same arithmetic run backwards.

Each correction takes a bite, and the shape of the result is a funnel rather than a rectangle. The figure above walks an invented example through all five steps: the count as measured, the count as people, the count as opportunities, what is still unrecovered, what a change would newly reach, and how many of those become sales.

Nothing in that example is a benchmark. The numbers are made up on purpose, because there is no defensible industry figure for a conversion rate on missed enquiries in your trade, in your region, this year. What there is instead is your own data and your own judgement, and the honest thing for a calculator to do with a box you cannot fill is to leave the answer blank.

Three of the inputs are forecasts rather than facts: how many callers already come back, how many of the rest a change would reach, and how many of those would buy. Presenting a single number built out of three guesses is false precision, however carefully the multiplication was done.

Give each of the three a cautious and an optimistic end and run the whole chain three times. The cautious column is the one to look at first, and the useful test is whether it still covers the cost. If it does, the decision is easy. If only the optimistic column does, you now know exactly which assumption the case rests on, and you can go and measure that one thing instead of arguing about the total.

Note that "cautious" for the natural recovery share is its high end, not its low one. More people coming back on their own leaves less for anything to recover. It is easy to get that backwards and end up with a low scenario that is quietly the rosiest of the three.

The missed-call recovery calculator does all of this in your browser. Nothing is pre-filled, every box you leave empty reads as unknown all the way down, and it prints the three scenarios side by side with the assumptions behind each one. It also names the edges rather than computing through them: a hundred percent recovery, a hundred percent conversion and a period with no missed calls each produce a sentence instead of a plausible figure.

If the case survives, the next question is what you would actually change. Most of the recovery a small business finds is in where a call goes when nobody is free rather than in any purchase, and that is worth mapping before you spend anything. If it turns out you do need more capacity on the phones, what a seat costs is the other half of the decision, and it belongs in the cost box rather than in a separate conversation.