Our method
Every figure on this site about cities, specialties and the options we compare names its source and the year it refers to, or says that it is a ClinicAll estimate. This page explains what a sourced figure is, how the calculator works, how we compare, how often we re-check and how to report an error.
What a sourced figure is
For each sourced figure, our data keeps its value, the date it was true, the source that publishes it and the day we consulted that source. The tables and the calculator read this data, and the source line under a figure names the publication and the year the figure refers to.
A source is a public page you can open yourself: INSEE, the French Assurance Maladie, the Luxembourg CNS, a vendor’s own website. When the number sits in a PDF or a spreadsheet, our data records the page or table where it is printed, so that it can be checked by hand.
The texts are written with the help of AI models, from a file of sourced facts specific to each page. Every figure written in a text is checked against our data by a second AI model that did not write it. Two automatic checks cover figures and wording: the site build fails if a published page shows a figure whose source line is still only a stand-in, or if it contains a phrase from our list of banned phrases, such as a promised result.
How the calculator works
The missed-calls calculator starts from five values: calls per day, the share of those calls that goes unanswered, the share of unanswered callers who would have booked, the value of a visit and working days per month. Under each starting value taken from our data, a line names its source; once you change it, the field reads “Your figure” and offers a way back to the starting value.
Working days start from an average month in France, Monday to Friday less public holidays. Together with the share that would have booked, they sit under “Assumptions”, which you can open to change them.
The calculator multiplies calls per day by the unanswered share: that gives the unanswered calls per day. Multiplied by the share that would have booked and by the working days, they give the missed bookings per month; multiplied again by the value of a visit, the revenue at stake per month. The result shows that amount first, then the missed bookings per month and the yearly amount, which is the monthly one times twelve. Under “How it adds up”, the same sum is written out with your values.
Without a visit value the calculator puts no amount on your behalf: it shows the missed bookings per month first and asks you to set the value of a visit. The result is an estimate built on the values in the fields, not a forecast of your revenue.
What we estimate
We found no public study that gives, per specialty, the number of calls a practice receives in a day, the share that goes unanswered or the share of those callers who would have booked. For these three values the calculator starts from ClinicAll estimates, whose source line links to the French version of this page. We plan to replace them with figures drawn from the calls ClinicAll handles.
These estimates do not describe your practice. In the calculator, replace them with your own figures: your phone system’s call log gives the number of calls and the share left unanswered. The result is recalculated from your values, and on the calculator page the link you copy keeps them.
The value of a visit is not an estimate: it comes from a published source, most often an Assurance Maladie tariff. Where no source prices a typical visit in a specialty, we use the nearest published figure, such as the fee for a related consultation, and its source line names that publication. For opticians and pharmacies we found no published figure: the field starts empty and you enter your own.
The table below groups the starting values that are ClinicAll estimates: one row per value, with the specialties it applies to.
| Calls per dayEvery specialty | 40 calls |
|---|---|
| Calls that go unansweredEvery specialty | 25% |
| Of those, would have bookedEvery specialty | 40% |
How we compare
A comparison sets ClinicAll beside another option, feature by feature. Each cell for the other option reads “Yes”, “No” or “Not stated”. A “Yes” or a “No” links to the vendor’s public page where we read it; for a category, such as a telephone answering service, the cells cite the public pages of named providers, listed under the table.
“Not stated” means the vendor does not publish the information: we neither infer it nor count it as a “No”. When more than half of the features are not stated, those rows leave the table and a sentence under it lists them, so that a column of “Not stated” does not read as a charge against the vendor. Prices are the ones the vendor displays, each linked to its page.
ClinicAll’s cells link to the page of this site that describes the feature. Where a feature has a limit, the cell states the limit instead of a plain “Yes”.
Every month, we reread the vendor pages each comparison cites and update the cells that have changed; the date of the last check sits under the table. A change on the vendor’s side therefore shows here only after the next check. Every comparison also says when the other option is the better fit.
Production figures
The usage figures on this site come from Hubert Vision, whose phone assistant has answered the practice’s line since 31 May 2026. They are drawn from the calls the assistant handled: inbound calls, distinct callers, appointments booked, new patients, the median call length and how calls spread over the days and hours of the week.
Their definition stays the same from one month to the next, so that the figures remain comparable, and each figure states the period it covers. The numbers of calls, callers, appointments and new patients are exact counts, never estimates. Only the median call length, to the nearest second, and the share of calls outside office hours, to the nearest percent, are rounded.
For the totals, the calculation runs on the server that holds these calls, and only the results leave it. For the spread by day and hour, only each call’s start time, direction and status leave it. No name, phone number or call content leaves the server, and nothing on this site identifies a patient.
When the figures are re-checked
Every quarter, we run a verification program that reopens every source in our data: it checks that each page still answers and that each figure is still printed on it. A figure in a PDF or a spreadsheet is re-checked by hand, at the place our data records. ClinicAll estimates have no outside source to reopen, so the program lists them to keep any from being forgotten. It identifies itself as “ClinicAll site data verification” and links to the French version of this page.
Every year, we add by hand the public and school holiday calendars for the year ahead, as the authorities publish them. Every month, we update the production figures of Hubert Vision and the spread of its calls by day and hour, and we re-check every comparison.
When a source changes, we correct the figure in our data. The tables and the calculator that show it then take the new value, and we correct by hand the texts that quote it.
Report an error
If a figure looks wrong, a source no longer answers or a fact is out of date, write to us at the address below and name the page. A vendor in a comparison can write the same way to have a cell corrected.
We check the source before correcting anything. A confirmed error is corrected in our data and in every text that quotes the figure.
See it answer a call for your practice
A 30-minute demo by video. We show the assistant live and look at how it would fit your practice, your software and your rules.
Book a demo