Software in French: what our 123 reviews demonstrate

What We Count

We frozen the catalogue data set at 28 September 2026, then counted its statutes for French. This is an analysis of our existing documentary evidence, not a new trial of all these products. The dates of the individual observations, between the 20 and the 27 September 2026, are kept in the downloadable data.

These percentages describe our catalogue and the coverage of its evidence. They don't measure the software market. In particular, 74,8 % unconfirmed cases does not mean that 74,8 % software is in English. Absence, imprecise or insufficiently exploited documentation in our file does not demonstrate the absence of a translation.

"In French" can mean four different things

A small business buys a tool to perform tasks, not to read its commercial home page. However, the same language may refer to the site of the vendor, the work interface, content intended for customers or support. Consolidating these surfaces produces too wide a promise.

Shopify explicitly distinguishes the language of the administration of the shop. This is the type of distinction sought in our survey: the fact that a merchant can present his site in French does not in itself establish the language of his management space. The same vigilance applies to a tool that generates a video or translates an email into French.

For an independent working alone, reading some English menus can be acceptable. For a team that regularly welcomes new employees, the language of settings, error messages and support can weigh more than that of the showcase. This difference is due to the need for use; it is not just a universal classification of products.

A reproducible method, with a visible blind angle

1. Population. 123 entries in the language file, including tools whose sheet remains excluded from indexing. The study does not reopen any of these pages.
2. Unit. A catalogue software, counted once for the French language. The other languages do not increase the total.
3. Classification. An unconfirmed marked file remains unconfirmed. Otherwise, the programme reviews successively the lists of available, partial and unavailable languages; an absence in these lists leads to unconfirmed.
4. Calculation. The number of entries of each status is divided by 123. The shares shall be rounded to one decimal place, except zero.
5. Traceability. Each line keeps the source vendor, the date of the statement and its note. The snapshot also indicates the source commit and the fingerprint of the file used.

The absence of partial cases does not prove that all documented interfaces are fully translated into all modules. It simply indicates that no entry in the statement was classified as "partial French". Mobile versions, add-ons and developments after the check dates may request a separate check.

Our main limitation is therefore the available documentation. A future study may reduce the 92 unknowns after further checks, but it will need to publish a new snapshot and explain the changes. We will not retrospectively replace the unknowns in this edition with conclusions found later.

Download 123 lines in CSV · Consult the instant JSON and its sources · See current language statement

Questions to ask before forming a team

Ask for proof of the precise route your colleagues will use. For a CRM: Create a business, change your step, find an error message and export a list. For a store: modify a product, process an order and include a refund. These are proposed audit scenarios, not tests that Logicelio claims to have performed.

In a demo, distinguish the settings of the personal account from those applied to all employees. Also ask if the help and support follow the language of the interface. Keep a copy of the vendor's answer with the relevant date and package: an isolated capture of a menu does not necessarily cover your entire use.

The Shopify review allows to continue on e-commerce; category CRM is used to build a selection according to the profession. Our methodology explains the difference between documentary evidence, external testimony and actual essay. This study does not add any user reviews or field measurements.