What should a small business track about its enquiries?

Most small businesses do track enquiries, and most of them track the wrong population. The log begins when someone enters a lead into a system, which happens after a human has already decided the enquiry is worth entering. Everything that failed before that point — the call nobody answered, the form that went to an unread mailbox, the message on a channel nobody watches — is absent by construction. The system reports accurately on the enquiries that survived and says nothing about the ones that did not.

Fixing this needs five fields and no software purchase. Arrival time, to the minute. The channel it arrived on. The source that produced it, where that is knowable. The time of the first response from a person, with automated acknowledgements not stopping the clock. And the outcome, with no response recorded as an outcome rather than as a blank. The last of those is the whole point: an enquiry that received nothing has to be a row, or the number that matters is unrepresentable.

Those five answer nearly every operational question a small business actually has. Where do the enquiries come from that turn into work, and where do they come from that do not. What time of day and day of week does the intake break down. Is the response problem general or concentrated in particular bands. Which channels are being monitored and which are theatre. Whether last quarter's change to the phones did anything. None of that requires a further field, and adding fields before these five are reliably populated is the standard way tracking projects collapse.

The reason to timestamp arrival separately from first response is that they measure different things and get conflated constantly. Arrival is a fact about the customer. First response is a fact about the business. The gap between them is the only measure that describes what the customer experienced, and businesses that record only when they acted are measuring their own diary. It also matters that an automatic acknowledgement does not stop the clock — otherwise the metric can be improved to near zero without any change to when a person actually engaged.

On outcome, the useful resolution is coarser than most systems offer. Won, lost, no response, out of scope. Elaborate stage models look thorough and get filled in unevenly by busy people, which produces a dataset nobody trusts and eventually nobody maintains. Four values entered reliably beat twelve entered when someone remembers, and the four are enough to compute everything above.

The last discipline is that the log should be reviewed on a schedule by someone with authority to change something. An enquiry log nobody reads is a database, and its main effect is to make people feel that the problem is being handled. Fifteen minutes a month, looking at the no-response rows and the longest response gaps, is what converts it from a record into a management instrument.