Key takeaways
- A useful dashboard has a dozen measures or fewer, each with a written definition, an owner and a trigger level that prompts action.
- Cover four areas: access, quality and safety, money, and people.
- For GP practices in England, the five 2026/27 demand metrics are a natural core for the access section.
- Match the review rhythm to the measure: some belong in the daily huddle, others in a monthly or quarterly meeting.
Why most dashboards fail
Most practice dashboards fail in one of two ways. Either they try to show everything the clinical and phone systems can export, so nobody can tell what matters, or they are built once for a meeting and never updated. A working dashboard is small, regular and tied to decisions. Every number on it should answer the question: if this moves, what will we do?
Start from the questions partners, owners and managers actually ask. Can patients get through? Are we safe? Are we paid for what we do? Is the team coping? Each question becomes a section with two or three measures, and anything that does not help answer one of them stays in the detailed reports.
Rule of thumb: if a measure has not triggered a conversation or a change in the last six months, either its trigger level is wrong or it does not belong on the dashboard.
Choosing the measures
For GP practices in England, the access section can start from the five demand metrics NHS England asks practices to collect for 2026/27: call waiting time between 8am and 10am, call waiting time across core hours, the share of clinically urgent requests seen on the same day, and the share of non-urgent requests seen within one and within two weeks[1]. Practices will be collecting the data anyway, and the measures map directly to what patients experience.
| Area | Measure | Source | Review |
|---|---|---|---|
| Access | Average call wait 8am to 10am | Cloud telephony report | Daily and weekly |
| Access | Calls abandoned as a share of calls offered | Cloud telephony report | Weekly |
| Access | Clinically urgent requests seen same day | Clinical system | Weekly |
| Access | Non-urgent requests seen within 14 days | Clinical system | Monthly |
| Access | Online requests waiting over one working day | Online consultation system | Daily |
| Quality | QOF points achieved against plan | QOF reporting service | Monthly |
| Quality | Significant events and complaints opened and closed | Practice log | Monthly |
| Experience | Friends and Family Test positive share | FFT returns | Monthly |
| Money | Income against budget, top three lines | Accounts or payroll | Monthly |
| People | Sickness absence and vacancies by role | Rota and HR records | Monthly |
| Efficiency | Missed appointments and slots refilled | Clinical system | Weekly |
Eleven measures is plenty. Fewer is better if the team is small. Resist adding measures that are interesting but do not lead to action, such as total calls received on their own, which rise and fall with demand and say little about whether patients were helped.
Missed appointments earn their place on almost every dashboard. In January 2026 alone, 1.3 million GP appointments in England were missed or unattended[2]. Tracking both the missed count and the share of those slots refilled shows whether reminders and cancellation lists are doing their job.
Define every measure in writing
Two people reading the same number differently is the fastest way to lose trust in a dashboard. For each measure, write a one-line definition, the exact data source and report name, who updates it, when, and the trigger levels that turn it amber or red. Keep these definitions on a second page of the dashboard so they travel with it.
Name it plainly
Call it what it is, such as average wait to answer between 8am and 10am, rather than an internal code.
Define numerator and denominator
For a share, state exactly what is counted on top and underneath, including what is excluded.
Name the source
Record the system and report, with any filters, so anyone can reproduce the figure.
Set triggers
Agree green, amber and red levels from your own recent history, not from national averages alone.
Assign an owner
One person updates it and one person, often the same, explains it at the review.
Use trends, not single numbers
A single month tells you very little; a trend tells you a lot. Show each measure as a short line over 12 months with the trigger bands shaded. National data shows why direction matters: the share of patients who found it easy to contact their GP practice by phone rose from 49.7% in 2024 to 52.9% in 2025 and 56.9% in 2026[3]. A practice that knows its own direction can tell partners whether changes are working long before the annual survey arrives.
Source: Ipsos (GP Patient Survey provider for NHS England), 2026[3]
Show the data
| Period | Value |
|---|---|
| 2024 | 49.7% |
| 2025 | 52.9% |
| 2026 | 56.9% |
Add one national or peer reference line where a fair one exists. Patient feedback is a good example: in June 2026, 92% of Friends and Family Test responses about general practice in England were positive and 4% negative[4]. A practice consistently well below that line has something to investigate; one consistently above it has something to protect.
Set the review rhythm
- Daily huddle, five minutes: yesterday's call wait from 8am to 10am, online requests still waiting, staff absence today.
- Weekly operations meeting, 20 minutes: abandoned calls, same-day urgent access, missed appointments and slots refilled.
- Monthly partners or owners meeting: the full dashboard, with exceptions explained and actions agreed.
- Quarterly review: trigger levels, definitions and whether each measure still earns its place.
Keep the format identical every time. People learn to read a dashboard by its layout; moving sections around each month forces them to start again. A single page with the same four sections in the same order works better than a long report with more detail.
Designing the page so it gets read
Layout matters more than software. A spreadsheet printed on one sheet of A4 that everyone reads beats a sophisticated business intelligence tool that only the manager opens. Whatever you use, the same few design choices make the difference.
- Put the four sections in the same order every month: access, quality and experience, money, people.
- Show the latest value in large type, the trend as a small line beside it, and the trigger colour behind it.
- Write one sentence under any amber or red measure explaining why it moved and what is being done.
- Use the same scale for related measures so the eye compares like with like.
- Avoid traffic lights without numbers; a red square on its own invites argument rather than action.
- Date-stamp the page and name the person who prepared it.
Test the page by handing it to a partner or the reception lead who has not seen it before and asking them what needs attention. If they cannot answer within a minute, simplify it.
Dental practices and private clinics
Dental and private clinic dashboards follow the same structure with different measures. Access becomes enquiry response time, time to first appointment and chair or room utilisation. Money includes treatment plan acceptance, revenue per clinical hour and, for NHS dental contracts, contract delivery against plan. People covers associate and hygienist availability and nurse vacancies.
- New patient enquiries answered within one working day.
- Days from enquiry to first appointment.
- Chair or room utilisation by clinician.
- Treatment plan acceptance rate.
- Failed appointments and short-notice cancellations.
- Recall compliance: patients seen within their recall interval.
Example: a fictional surgery's first dashboard
Example: Greenfield Surgery, a fictional practice with 10,500 patients, starts with nine measures on one page. In the first month, the average wait to answer between 8am and 10am is nine minutes against an amber trigger of six, and 60 online requests a week wait more than a working day. The partners agree two actions: move one receptionist's admin hour out of the morning peak and add a late-afternoon online request review.
Example: three months later the morning wait is under six minutes and requests waiting over a day are in single figures. The team drops one measure that never moved and adds missed appointments refilled, because reminders have become the next priority. The dashboard has changed decisions twice in a quarter, which is the clearest sign it is working.
Common traps
- Measuring activity instead of outcomes: calls answered tells you less than calls abandoned or the wait to answer.
- Setting triggers from national averages that do not fit your list or your building.
- Changing a definition without marking the break in the trend line.
- Letting the dashboard become a blame tool; it works best when the people who own the measures also propose the fixes.
Keeping it alive
- Automate exports where you can, so updating takes minutes not hours.
- Publish on the same day every month.
- Show actions agreed last time beside the numbers they were meant to move.
- Retire measures that no longer lead to action.
- Share a version with the whole team, not just the partners.
Frequently asked questions
How many measures should a practice KPI dashboard have?
Around eight to twelve. Each should have a definition, an owner and a trigger level, and each should lead to a decision when it moves.
Which access measures should a GP practice track in 2026/27?
Start with the five national demand metrics: call waits from 8am to 10am and across core hours, clinically urgent requests seen the same day, and non-urgent requests seen within one and two weeks.
How often should the dashboard be reviewed?
Some measures daily or weekly in operations huddles, the full dashboard monthly with partners or owners, and the measures themselves quarterly.
References
NHS England (2026). Changes to the GP Contract in 2026/27 (opens in a new tab) ↑
Wessex Local Medical Committees (from NHS appointments data) (2026). Appointment Overview January 2026 (national) (opens in a new tab) ↑
Ipsos (GP Patient Survey provider for NHS England) (2026). 2026 GP Patient Survey results released (opens in a new tab) ↑
NHS England (2026). Friends and Family Test data collection infographic, June 2026 (opens in a new tab) ↑