Research
The 14-minute response window
How long it takes a patient enquiry to get a first clinician response is the metric we obsess over. Past 14 minutes, the patient is half as likely to engage.
There’s one metric we look at every Monday morning before anything else: the median time from a patient enquiry landing in our system to a first clinician response.
For most healthcare businesses this would not be the headline number. Patient outcomes, prescribing rates, satisfaction surveys, complaint counts, all of those matter and we track them. But the response-time number sits above all of them in importance for one reason: it’s the first thing in the funnel, and what happens here propagates through everything downstream.
The shape of the curve
We’ve done a retrospective analysis on roughly 140,000 first enquiries across two years. The question we asked: given a patient who submits an enquiry, what fraction complete their journey (define “journey” as completing a first consult within 7 days)?
The answer, simplified:
| Response time | Completion rate |
|---|---|
| Within 5 min | ~88% |
| 5–14 min | ~71% |
| 14–30 min | ~36% |
| 30–60 min | ~22% |
| 60+ min | <12% |
The drop between 14 minutes and 30 minutes is the steepest part of the curve. Patients who get a response within 14 minutes are roughly twice as likely to complete as those getting a response at 30. Beyond an hour, you’ve effectively lost them.
Why 14 minutes specifically
There’s no clinical magic in the number. It’s an empirical observation about how long a person stays in a particular state of attention after submitting an online form for healthcare.
Plausible drivers:
- Patient context decay. The combination of motivation and attention that produced the enquiry doesn’t persist. Other things in their life intervene. The medical issue stops feeling acute.
- The “I’ll do it later” trap. People who reach out about healthcare and don’t get a response within their attention window often move on, sometimes for weeks, sometimes forever. The condition usually doesn’t.
- Trust signalling. A fast response signals competence. A slow one signals “we may be like the rest of healthcare.” First impressions calibrate expectations for the entire relationship.
What we do operationally
Defending the 14-minute window is not free. It requires:
- Staffing for peak load, not average. Patient enquiries are not uniformly distributed. They peak at lunchtime, after work, and in the first hour of weekend mornings. Staffing models that target average response time fail at peak time precisely when the most enquiries come in.
- Accepting capacity that’s unused. Some hours, the response queue is empty. Clinicians are paid for that hour anyway. The alternative is a queue, which we don’t tolerate.
- A real-time queue with no triage gatekeeping. Most online clinics route through a non-clinical triage step first. We mostly don’t. The first response a patient gets is from a clinician with prescribing authority.
- Async messaging, not phone-tag. Phone calls require synchronous availability. Async messaging lets clinicians work through queues without missed-call ping-pong.
The trade-off we accept
The cost of all this is straightforward: lower utilisation of clinician hours than a sequential-queue model would produce. We accept that cost because:
- The completion-rate uplift more than pays for it.
- Patients who complete their first consult quickly are also the patients most likely to stay in care across time, which is the metric that actually matters.
- Reducing wait time reduces the number of patients who give up on the system entirely.
What this isn’t
We are not arguing that all healthcare needs 14-minute responses. Most of healthcare is fine on slower cadences. Emergency care needs a different model entirely. Specialist care can run on weeks of lead time.
But for outpatient access — the part of the healthcare journey we’re in — the cost of slow first response is too high. Specifically, it’s a cost that mostly falls on the patient, who tends to disengage rather than complain.
What we’d tell other operators
If you run a clinical service that depends on patient self-referral:
- Measure the time from enquiry to first response, not the time to first booking. First response is the leading indicator. Booking is the lagging one.
- Plot the completion curve. Even at small scale, the shape will tell you where your inflection point is. Don’t assume it’s the same as ours.
- Staff for the curve, not the average. Average response time is a misleading number. P50 (median) is OK. P90 is better.
We track many things. We obsess over this one because everything else flows from it.
This is one of five patterns we covered in one million consults: the five things that surprised us — the others are equally counter-intuitive.