Your practice
The cost of a patient you cannot fit is what it really costs you: lost goodwill, a comped visit, or overtime pay. If you would simply lose the visit, put a small goodwill figure here.
Your no-shows
One row per day. Booked is how many appointments were on the schedule; showed is how many turned up. At least 10 days and 60 appointments.
Enter your numbers, then click Find the best plan.
Result
Every overbooking level
Expected per day. "Over capacity" means more patients turn up than you have slots. The best row is highlighted by the bar.
| Extra | Booked | Visits done | Patients over | Chance anyone over | Chance a slot is empty | Profit / day |
|---|
How sensitive is the answer?
Forecast check
How it is calculated
Model: with B appointments booked and a show rate p, the number of patients who turn up is binomial. Because bad days cluster, the tool uses a beta-binomial, whose spread is larger by the factor 1 + (B − 1) × ρ, where ρ is the day-to-day correlation, estimated from your history by the dispersion of daily show rates around the overall rate (clamped to 0 to 0.5). Profit: for each plan it sums over every possible number of show-ups s: margin × min(s, slots) minus the cost of each patient beyond capacity (in overtime mode the margin is earned on every patient seen). No simulation: the distribution is exact. Uncertainty: the show rate gets a Beta posterior from your counts, with the evidence divided by the design effect 1 + (n − 1) × ρ so clustered data is not over-trusted, and each plan's profit is averaged over a fixed grid of 400 show rates. The 80% range and the best plan at each end are reported. Forecast check: the model is fitted on the first 75% of days and asked for the central 80% range of show-ups on the last 25%; the share of days landing inside should be near 80%. The same days are scored under independence to show whether clustering matters.
Not modelled: patients who cancel early and free a slot, different no-show rates by appointment type or weekday (run separate histories), same-day walk-ins, patients who are harder to rebook, and professional or regulatory rules on overbooking in your field. Unverified: checked against independent reference calculations and simulated histories with known rates in unit tests, not against a real practice's results or a commercial scheduling product.
Common questions
How many extra appointments should I book?
It depends on two things: how often booked patients fail to show, and what an over-full day costs you compared with an empty slot earns. Book one more patient only while the extra expected revenue from that booking is larger than the expected cost of the days it leaves over capacity. This tool computes that exactly from the full distribution of how many patients turn up, and shows expected profit for every number of extra bookings so you can see how flat or steep the top is.
What no-show rate do I need to make overbooking worthwhile?
There is no universal threshold, because it also depends on what an overflow patient costs you. As a rough guide, if the cost of fitting or turning away one extra patient is similar to the margin on a visit, a practice with about 5% no-shows gains little, around 10 to 15% usually justifies one extra booking per day, and 25% or more can justify several. Enter your own numbers: the answer is computed, not assumed.
Why does it matter that bad days cluster?
No-shows are not independent. A storm, a flu week or a holiday makes many patients drop out on the same day. The same average no-show rate then produces more very empty days and more very full ones than independent patients would. The tool measures this from your history as a day-to-day correlation and uses a beta-binomial distribution, which widens the spread of show-ups and makes the recommended overbooking more cautious.
What does the no-show range mean?
With a limited history your true show rate is uncertain. The tool gives an 80% range for it and averages the profit of each plan across that range, rather than trusting a single point estimate. Ten days of data gives a wide range; a few months gives a narrow one. The best plan at the low and high end of the range is shown so you can see whether the decision is sensitive to it.
What is the difference between bump and overtime?
If a patient arrives and there is no slot, you either send them away or rebook them, losing the visit and paying a goodwill cost (bump), or you see them after hours at extra cost and still collect the revenue (overtime). The tool supports both. Use bump for a hard-capacity service such as an imaging suite, and overtime for a clinic that can stretch the day.
How can I test this safely?
Start with the plan the tool recommends at its probability cap, for example no more than a one in five chance that anyone is over capacity, and track booked, showed and overflow for a few weeks. Paste the new history back in and re-run. Overbooking changes who you can see, so also check the policy against your professional rules and patient-safety obligations before using it.
Is my patient data uploaded anywhere?
No. You only need daily counts of booked and showed, never patient names. The file is read and analysed in this browser tab. Nothing is sent to a server, there are no accounts or cookies, and the tool never stores your data. It works offline once the page has loaded.