An operations manager is asked what the enquiries are about. She says a bit of everything, mostly complicated things that need a person.
Two weeks of tallying later, six categories account for well over half the contact, and five of the six have a documented answer that exists on the website already.
Nobody was wrong. They were describing the enquiries they remembered, which are the difficult ones, rather than the enquiries that consumed the time.
Step 1: define a small set of reason categories
Eight to twelve, drawn from a first day of listening. Too many categories and nobody tags consistently. Too few and the result tells you nothing actionable.
Step 2: tag every contact for a fortnight
Every channel, including telephone, email, forms, chat and anything arriving through social. Missing a channel produces a confident and wrong conclusion.
Step 3: capture the time of day and day of week
The distribution matters as much as the volume. Most businesses discover a significant share arriving when nobody is working, which is the pattern in why escape room enquiries arrive at midnight.
Step 4: record whether an answer already existed
For each contact, note whether the answer was documented somewhere the customer could have found. This column is what turns the audit into a plan.
Step 5: record the handling time honestly
Including the interruption cost rather than just the reply. A two minute question that breaks concentration costs considerably more than two minutes.
Step 6: separate first contacts from follow ups
A customer contacting four times about one thing is a different problem from four customers contacting once. The distinction changes the fix entirely.
Step 7: identify the questions that arrive incomplete
Enquiries that cannot be answered without asking something first are where most of the exchanges come from, which is the ground in glamping enquiries and what is actually included.
Step 8: find the questions people ask before they commit
Practical, anxious, specific, and rarely published. These are the ones that cost bookings, as described in claustrophobia and locked in questions.
Step 9: rank by volume times handling time
Not by volume alone. A medium frequency question that takes twenty minutes each time may be a bigger prize than the most common one.
Step 10: check where the answers currently live
Website, a document, one person's head, or nowhere. Answers that live in a person's head are a risk as well as a cost.
Step 11: write the top ten answers properly once
Plain, complete and reusable. Most of the return from an audit comes from this step rather than from any tool.
Step 12: publish them where the questions are asked
On the relevant pages, not on a general help page. An answer that exists but sits three clicks away is not being found by anybody.
Step 13: automate the categories that are pure information
Hours, coverage, price ranges, availability, process and status. These need no judgement and they are the bulk of the volume.
Step 14: repeat the audit in three months
To check what moved and what appeared. The list changes with the business, and a single audit is a snapshot rather than a system. The practical comparison of who should handle what is set out in website receptionist vs phone receptionist. See features, start at try, and the sector view sits under small business.