The call data you should be keeping

There's a number about your own business you almost certainly can't produce: how many people called last month, reached voicemail, and hung up without saying anything.

It isn't carelessness. Nothing in an ordinary phone setup records it. Voicemail keeps the people who left a message. The phone bill keeps minutes. Your cell holds a call log for a while and then rolls it off. The callers who tried and gave up leave no trace anywhere — which means the single most valuable thing about your phone is the thing you have no record of.

Six fields, that's the whole system

Per call:

When — date, time, and day of week.
Who — the calling number, or at minimum the area code.
What happened — answered, went to voicemail, or abandoned.
How long — duration.
What they wanted — one sentence.
Outcome — message taken, transferred, booked, quoted, nothing.

No CRM required. A spreadsheet does this. What makes it useful isn't the sophistication, it's the completeness: every call, including the ones nobody picked up.

What four weeks of it tells you

When you're unreachable. Missed calls are never spread evenly. They pile into the lunch hour, the stretch after five, Saturday morning, and the forty minutes you spend driving between jobs. That's a coverage problem with a shape, and you cannot see the shape without writing it down.

What people actually call about. The usual finding is that half the calls are the same three questions. Those three questions belong on your website, or in the hands of whatever answers the phone — and answering them there frees up the calls that are worth your time.

Whether your advertising works. Not clicks. Calls. A campaign that produces calls from a new area code is doing something. One that produces traffic and no calls is a hobby.

Repeat callers who never got called back. The same number three times in two days is a person actively trying to give you money, and they will stop.

The calls you don't need to record are the answered ones

Here's the part that inverts how most people approach this. You already have a record of every call that went well. The customer is standing in front of you, the job is on the calendar, the invoice exists. Answered calls are self-reporting.

Missed calls are invisible by construction. Nobody tells you about them, nothing bills for them, and the caller's next move is to dial the next name on the list — which happens in under a minute and generates no evidence at all. Whatever you believe your missed-call rate is, you formed that belief entirely from the calls you answered. It is the one metric where your intuition is guaranteed to be wrong in the same direction every time.

That's also why the pattern matters more than the total. A missed call is a data point about when your business is unreachable, and unreachable time clusters. Once you can point at the cluster, the fix is usually dull and cheap: somebody covering the phone over lunch, a forward rule, or something that answers and takes a proper message so the caller stops dialing down the list.

Which is the same reason the measuring problem and the answering problem turn out to be one problem. Anything that answers every call produces the record as a by-product — timestamp, number, reason, outcome, on calls you would otherwise never have known happened. You don't build the data set separately. You get it by not missing calls.

Two cautions

Metadata is not recording. Keeping who called, when, and what they wanted is a different thing from recording audio. Recording law varies by state, and some require every party on the line to consent — so if you're going to record, get advice for your state and announce it at the start of the call. Notes and timestamps carry none of that weight.

Keep less, for less long. Phone numbers are personal information, and a year of call metadata is plenty for spotting patterns. Treat the file the way you'd treat your customer list, because functionally that's what it is: don't email it around, don't leave it in a shared folder, and delete the old years.

What to do

Start counting this month, even crudely. A notebook by the phone beats nothing.
After four weeks, break the missed calls down by hour and by day of week. Look for the cluster.
Write down the three questions that make up most of your calls, and put the answers somewhere a caller can reach without you.
Fix the cluster. That's the whole return on the exercise.

The first month is uncomfortable, because the number of people who called and gave up is always higher than the guess. But it's one of the few things about a small business you can improve with a scheduling change and no additional spending — and you can't improve it while it's invisible.

If you want the phone measured and covered at the same time, tell me how yours is set up now.