How to measure your real response time
Measure response time the way a patient experiences it: from the moment a call, form, text or chat arrives to the moment a person offers a real appointment time. Combine your phone and form data with a month of planned test contacts at different hours, report the median and the worst case for each channel, and repeat it monthly.
Decide what you are measuring
"We answer pretty quickly" is an impression, not a measurement. Pick a small set of numbers and define them the same way every month.
For phones:
- Answer rate: calls answered by a person, divided by all calls, by hour of day
- Time to answer: rings or seconds until a person picks up
- Abandoned calls: callers who hung up on hold or in the menu
- Callback time: from a voicemail or missed call to the return call
For forms, texts and chat:
- Time to first human contact
- Time to a specific appointment time offered
For everything: how many requests became booked visits.
Report the median and the single longest time, not the average. One request that waited two days disappears into an average and is exactly the one you need to see. Keep office hours and after hours separate, because they answer different questions.
Use the data you already have
Most of the raw material already exists. It is just scattered.
- Phone system records. Your phone provider or office system can usually export every inbound call with time, caller number and outcome.
- Form submissions. Your website or form tool timestamps each one. The missing half is when someone first responded, so ask staff to log that in your request queue or practice management system.
- Texting platform. Inbound and reply times are usually recorded for you.
- Chat transcripts. Start time, handoff time and outcome.
- Your EHR. The time each appointment was created, which tells you when a request became a booking.
If you want to know which marketing source produced each call, add call tracking; see call tracking without losing your number. Keep this kind of operational data inside your own systems, and keep patient details out of marketing analytics tools.
Does ChatGPT recommend your practice?
The Surge Score measures five things about your practice website: how often ChatGPT, Gemini and Perplexity recommend you, your PageSpeed, your structured data, your indexing and your reviews, each with its evidence and date. Free.
Plan a month of test contacts
System data shows volume. Test contacts show the patient's experience, including the parts no system records, such as whether the person who called back offered a time.
Schedule twelve contacts over four weeks, spread across channels and awkward hours:
- Monday 8:05 am, phone call
- Tuesday 12:15 pm, web form
- Wednesday 4:50 pm, text
- Thursday 10:30 am, website chat
- Friday 4:45 pm, phone call
- Saturday 10 am, web form
- Sunday 9 pm, web form
- Monday 7:45 am, phone call before opening
- Tuesday 12:30 pm, text
- Wednesday 2 pm, web form
- Thursday 5:30 pm, phone call after closing
- The evening before a holiday, any channel
Use a personal phone and email with an obviously fake name and a routine new patient request. Tell the team in general terms that testing is happening this month, cancel any test appointment the same day, and never test clinical or on call lines without arranging it with the clinician first.
Keep a simple log
A shared spreadsheet is enough. One row per contact, with these columns:
- Date and time sent
- Channel
- Automatic reply received, and what it promised
- Time of first human contact, and by which channel
- Time a specific appointment was offered
- Who responded
- Notes: anything confusing, rude, impressive or wrong
Read it in this order. First, the worst result in each channel, because that is where a patient was most likely lost. Second, the pattern by hour, which usually points to lunch, the end of the day or the weekend. Third, the gap between first contact and an offered time. A fast callback that ends with "let me check and call you back" is still slow from the patient's side. For why that gap matters, see speed to lead.
Make it a monthly routine
One measurement changes little. A monthly page that the practice owner reads changes a lot.
- For each channel: requests received, median response, longest response, and how many met your target
- The same figures from last month, side by side
- Requests that became booked visits
- One fix chosen for next month, with an owner
Put it next to the other numbers you watch, such as new patients and visits; see the five numbers a practice owner should watch.
Share results with the team as a system problem, not a blame exercise. A slow lunch hour is usually a staffing pattern, not a person. Retest after every change, so you know whether the fix worked.
Traps that make response times look better than they are
Watch for these when you read reports:
- Counting automatic replies as responses. Only a person, or a system that can book, counts.
- Phone reports that call voicemail answered. Some systems record a call as answered when the voicemail picked up. Check how yours defines it.
- Ignoring menu hang ups. Callers who quit inside the phone tree may never appear as missed calls.
- Testing only at convenient times. Tuesday at 10 am flatters everyone.
- Staff knowing the exact test times. Share that testing will happen, not when.
- Counting a voicemail left by staff as contact. It is an attempt, not a conversation.
- Time zones and clocks. Exports from different systems may use different time zones, which can shift results by hours.
If the numbers look too good, recheck one of them by hand. For the dollar value of what the gaps cost, see what missed calls cost a practice.
Checklist
- Each channel has defined metrics measured the same way every month
- Phone, form, text and chat data are pulled into one place
- Twelve test contacts are scheduled across channels and awkward hours
- Results report the median and the longest time, not the average
- Office hours and after hours are reported separately
- One fix with one owner is chosen each month and retested
- You have confirmed how your phone system defines an answered call
Frequently asked
What is the most important response time number?
The time from a new patient's first contact to being offered a specific appointment, and the worst case for it each month. That single figure captures most of what decides whether a patient books with you or with someone else.
Is it fair to mystery shop our own front desk?
Yes, if you are open about it. Tell staff that response testing happens every month, use the results to fix systems rather than blame people, and cancel any test appointments the same day. Framed that way, the results become an argument for the coverage your team needs, such as help at lunch.
How often should we measure response time?
Monthly, with a small set of test contacts plus your system data. Measure again after any change in staffing, hours, phone system or tools, since that is when response times usually shift.
Book a strategy call with the team.
Twenty minutes. We'll walk through the specific opportunities in your market and what a MedPracticeGrowth engagement would look like for your practice.