Key takeaways
- Five reports cover most decisions: calls by half hour, waits, abandonment, call-backs and calls per staff member.
- NHS England highlights telephony management information, such as call volume and length per staff member, as a tool for capacity planning[1].
- Read calls together with online requests: requests per 1,000 patients have risen sharply since online access became mandatory in core hours[2].
- Turn each weekly review into one decision, written down with the date you will check it.
- Use the data to fix systems and rotas, not to rank individual receptionists.
Why most telephony data goes unread
Cloud phone systems produce more information than most practice managers have time to read. NHS England's access recovery plan points out that digital telephony gives detailed management information, such as call volume and the length of calls for each member of staff, to support capacity planning and quality improvement[1]. In practice, the dashboard often stays closed because nobody has decided which few numbers matter.
The approach in this guide is deliberately narrow. Pull the same five reports at the same time every week, spend 30 minutes on them, and leave with one written decision. Over a quarter, that adds up to a dozen concrete changes, each tested against the following weeks' data. The same plan also notes that Leeds practices used real-time feedback on call length and waiting times to match staff to demand[1].
The five reports to pull every week
- Calls offered by half hour and weekday. This is your demand curve. It shows when patients ring, not when you would like them to.
- Average and longest wait by half hour. The average tells you about the typical caller; the longest wait tells you about the worst experience that week.
- Abandoned calls by half hour, using your agreed definition. Exclude very short hang-ups and count call-back requests as handled.
- Call-backs requested and completed. A gap between the two is a broken promise to patients.
- Calls handled and average call length per staff member. Use this to spot training needs and uneven workload, not to set individual targets.
If your system can export raw call records, keep a simple spreadsheet with one row per week for each of these, so you can see trends without logging in to the supplier's portal.
A 30-minute weekly routine
Minutes 0–5: check the week was normal
Note anything unusual: bank holidays, sickness, a clinical system outage, a local outbreak. Unusual weeks still teach you something, but do not let them drive permanent changes.
Minutes 5–15: find the worst half hours
Look at calls offered, wait and abandonment together. Mark the three half hours with the longest waits or highest abandonment.
Minutes 15–20: check staffing in those slots
Compare with how many people were logged in to the queue, not how many were rostered. The gap between the two is often the answer.
Minutes 20–25: check call-backs and online requests
Were all call-backs completed the same day? Did online requests rise or fall in the same week?
Minutes 25–30: write one decision
For example, “Add one person to the phone queue from 8.00 to 9.00 on Tuesdays for four weeks, then review.” Put it in the practice meeting notes with a review date.
From pattern to decision
| What you see | Likely cause | What to try |
|---|---|---|
| Long waits and high abandonment at 8.00–9.00 | Too few people on phones, or all capacity released at 8am | Fix the phone team for the first hour; release capacity in waves |
| Waits rise at 13.00–14.00 | Lunch breaks taken at the same time | Stagger breaks; keep at least one person on the queue |
| Many call-backs not completed by close | Call-back offered too late in the day | Set a cut-off based on estimated wait |
| One person's average call length far above others | Taking harder calls, or needs support | Listen to sample calls together before drawing conclusions |
| Hang-ups during the menu | Menu too long or confusing | Cut options; put the most chosen first |
| Calls fall but online requests rise | Patients switching route | Check online requests are being handled within the promised time |
Example: a fictional practice saw waits climb every Thursday after 14.00. The calls offered were normal, but staff logged in dropped from four to two because Thursday afternoon was the scanning and coding slot. Moving that work to Friday morning, which was quieter on the phones, cut the Thursday wait on the following weeks' reports. The data did not tell the practice what to do; it told it where to look.
Read calls and online requests together
Phone data on its own tells half the story. Since October 2025, practices in England have had to keep online consultation tools open throughout core hours. The Care Quality Commission reports that practices received 9,617,628 online requests in March 2026, 69% more than the 5,690,082 received in August 2025, before the requirement, and that there were 153 online requests per 1,000 patients in March 2026[2].
Source: Care Quality Commission, 2026[2]
Show the data
| Item | Value |
|---|---|
| August 2025 | 5,690,082 |
| March 2026 | 9,617,628 |
Put your practice's online request count on the same weekly sheet as calls offered. If both rise together, demand has grown. If calls fall as online requests rise, patients are switching channel, and the next question is whether online requests are being answered within the time you promise. If calls rise when online requests fall, check whether the online form was switched off or broken that week.
Monthly and quarterly views
Weekly reviews catch operational problems. Monthly and quarterly views show seasonal patterns and the effect of bigger changes. NHS England's own statistics note that call volumes are affected by the number of days in a month, public holidays and weekend days[3], so compare per working day rather than monthly totals.
- Monthly: calls offered per working day, abandonment, average wait and call-back completion, with a one-line note on anything unusual.
- Quarterly: the same figures compared with the same quarter last year, plus the share of requests arriving online.
- Before winter: last year's October to January calls by week, to plan flu season and Christmas cover.
- After any change: the four weeks before and after, using the same definitions.
For scale, around 1.55 million standard GP appointments a working day were delivered in England in August 2026[4]. Your own appointments per working day, set beside calls per working day, gives a simple ratio that shows whether phone demand is growing faster than capacity.
Presenting call data to partners
Partners rarely want a full dashboard. They want to know whether access is getting better, where the pressure is and what the practice is doing about it. A single page each month, presented in the same format, builds trust in the numbers and makes decisions faster.
- One line on the headline: “Average wait at 8am fell this month; abandonment at 8am is down; online requests are up.”
- A small chart of calls offered and abandoned per working day over the last six months.
- The three worst half hours this month and what is being done about each.
- The decisions made in weekly reviews, with whether each one worked.
- Any request for partner input, such as agreeing a change to appointment release times.
Keep explanations of definitions in an appendix the first time and refer to it afterwards. When partners trust that the figure means the same thing every month, discussion moves from questioning the number to deciding what to change.
Questions to ask your supplier about reporting
- Can we export every call record with start time, queue time, outcome and the menu option chosen?
- How does the system define an abandoned call, and can we exclude very short hang-ups?
- Are call-back requests shown separately from abandoned calls?
- Can we see unique callers as well as total calls, so repeat dialling is visible?
- Can reports be scheduled to arrive by email every Monday morning?
- How long are call records kept, and can we set a shorter period for exports that contain phone numbers?
If the answer to the first question is no, the other reports will always be limited. Raw call records are the foundation of every useful analysis described here.
Use the data fairly
- Share the weekly sheet with the reception team. People who answer the phones often know why a number moved.
- Do not publish individual call counts on a wallboard. Show the queue, not the people.
- Remember that call logs contain phone numbers and sometimes names, which are personal data. Limit who can export them and delete old exports.
- Keep definitions stable. If you change how abandonment is counted, note it on the sheet so trends are not misread.
- Celebrate improvements. A visible drop in the 8am wait after a rota change builds support for the next change.
For the definitions behind these reports, see what abandoned call rate is normal. To turn twelve weeks of data into a forecast, read forecasting phone demand, and for converting a forecast into a rota, how many receptionists you need on the phones.
Frequently asked questions
How often should a practice manager review call data?
A short weekly review of the same five reports works for most practices, with a monthly summary for partners and a quarterly comparison with the same period last year.
Should we use call data in staff appraisals?
Use it carefully. Call counts and lengths depend on which calls a person happens to take. Listening to sample calls together is more useful for development than ranking people by numbers.
Our supplier's reports do not show half-hour data. What can we do?
Ask for a raw call record export, which usually includes start time, wait and outcome. A simple spreadsheet can group these into half hours.
What should we share with the PCN or ICB?
Aggregate figures such as calls per working day, average wait and abandonment are usually enough. Avoid sharing raw call logs, which contain personal data.
References
NHS England and DHSC (2023). Delivery plan for recovering access to primary care (opens in a new tab) ↑
Care Quality Commission (2026). State of Care 2025/26: capacity in primary care (opens in a new tab) ↑
NHS England (2026). Statistical Note: Integrated Urgent Care Aggregate Data Collection (IUC ADC) July 2026 (opens in a new tab) ↑
British Medical Association (2026). Pressures in general practice data analysis (opens in a new tab) ↑