Should a practice add overflow phone cover?

The reason this decision stays open for years is that both sides of it are unmeasured, but only one side knows it. The cost is exact and arrives monthly: overflow cover is priced per call or per minute, and a supplier will quote it in an afternoon. The benefit is a count of calls that nobody in the practice has ever seen, because the calls in question are the ones that did not reach anybody. So the decision is made by comparing a precise number against an absent one, and an absent number loses every time.

What makes the absence structural rather than sloppy is where the record gets written. Most business phone configurations create the call record on answer. A caller who hangs up while ringing, who reaches a full queue, or who hits voicemail outside hours does not produce a row in the report the practice reviews. Nobody is hiding it. The system was never asked to write it down, and the practice's honest impression — that the phone is answered most of the time — is an accurate summary of every call it has data for.

There is a measurement that gets around this entirely, and its usefulness comes from being taken from the outside. The share of inbound calls that a person actually answers can be established by calling the practice's own published numbers across a spread of times, inside and outside stated hours, and recording what happens: a person, a queue, a voicemail, or nothing. Voicemail counts as unanswered, because a patient choosing between three practices treats it that way. No access, no credentials, and no cooperation from the phone system are required, which is what makes it usable as a baseline before any purchase and as an audit afterwards.

Run that once and the decision usually answers itself, because the result is rarely ambiguous. Practices tend to discover either that cover inside hours is adequate and the loss is concentrated in a narrow band — lunch, the first hour, the last hour — or that the out-of-hours picture is a great deal worse than anybody assumed. The first finding is a rota problem and does not need a supplier. The second is what overflow cover is actually for, and it is a much smaller purchase than the always-on contract usually quoted.

The trap on the other side is buying cover and measuring nothing afterwards. Per-call pricing means the invoice scales with the volume you never counted, and the only way to know whether the service is answering calls that would have converted, or answering calls that would have called back anyway, is to keep taking the same outside measurement after it is live. A supplier who is answering will not object to being measured; a supplier who is queueing will.

One boundary worth setting in the contract rather than discovering later: what the answering party is permitted to say and record. An overflow service that takes a name, a number and a callback window is administrative. One that takes clinical detail is handling protected information, and that is a different agreement with a different vendor obligation behind it.