Minitel

A French state network that from the early 1980s handed households a free terminal in place of the paper phone book and then grew thousands of paid services on top of it, from chat lines to banks, all billed straight to the telephone account. A decade before the mass internet it was used by as many as 9 million terminals; it was shut down in 2012.

In brief

Minitel was a French state network that, a decade before the mass internet, handed millions of families a free terminal in place of the paper phone book and then grew a whole parallel economy on top of it: chat, shopping, banks, tickets, news, all billed straight to the telephone account. At its peak (1999–2000) the network was used by about 9 million terminals and 25 million people, close to half the population of France, and its most profitable and most talked-about segment was the paid pink chat lines, the messageries roses. The state shut the network down on June 30, 2012, exactly 30 years after the commercial launch, when somewhere between ~670,000 and ~810,000 terminals were still in circulation, depending on the estimate.

How it started (the Nora–Minc report, DGT, Théry)

In 1978 Simon Nora and Alain Minc wrote the report "L'informatisation de la société" at the request of President Valéry Giscard d'Estaing, and it was there that the term telematics (télématique), the merging of telecommunications and computing, first appeared. The report proposed not a closed system but the digitization of the telephone network and an open platform for services on top of it.

The idea was carried out by Gérard Théry, head of the DGT from 1974 to 1981: in September 1977 he commissioned a report on videotex technologies, and in November 1978 the Council of Ministers approved a plan of six projects, the videotex directory among them. On June 2, 1979 Théry formally opened the Télétel and Electronic Directory projects, and Giscard d'Estaing approved the program after a demonstration at the Élysée Palace. Taking it to scale would fall to someone else: in May 1981 the socialist Mitterrand became president, and the mass distribution of terminals began in 1983 under the new government. Decades later Théry assessed the invention without any triumph, saying that Minitel had not developed the way they had hoped, and in March 1986 he spoke of difficulties that were far more apparent than real.

Year-by-year timeline

1978: the Nora–Minc report introduces the term telematics and makes the case for an open platform on top of the telephone network fact
1979-06-02: Théry, as head of the DGT, formally opens the Télétel and Electronic Directory projects fact
1980: two parallel experiments, an electronic directory in Saint-Malo and Ille-et-Vilaine (55 volunteers, July 1980) and "Télétel 3V" in Vélizy with the neighboring communes (~2,500 households, ~200 providers) fact
1981-05: 4,000 terminals of the new type are distributed across the Ille-et-Vilaine department fact
1982: commercial launch across France, by then under President Mitterrand fact
1983 (end): 120,000 terminals in service; the mass free distribution begins fact
1984: launch of Kiosque (code 3615 and its neighbors), paid third-party services with revenue sharing through the telephone bill fact
1985–1986: messageries and games already account for 42% of network traffic; by the beginning of 1986 there are 1.4 million terminals (12% of households); on 1986-12-02 la Coordination, the student body opposing the Devaquet law, launches its own Minitel service for the strike committees, a supporting channel for several tens of thousands of people rather than the main one (television was still that) fact
1987–1988: Minitel Rose brings in 1 billion francs of revenue (a third of it to France Télécom); 3 million terminals; Kiosque providers receive about two thirds of what the customer pays fact
1990s (1999–2000): peak usage, about 9 million terminals and 25 million users fact
1998: network revenue of €832 million, of which €521 million is paid out to third-party service providers; France meanwhile lags on the internet, used by only 2% of the population fact
2012-06-30: France Télécom switches Minitel off; between ~670,000 (Al Jazeera) and ~810,000 (Wikipedia) terminals remain in circulation estimate

Lesser-known but significant facts

At the end of 1984 the PTT proposed keeping a history of the last pages visited and met a rebuff so sharp that around 3,000 terminals were returned in protest, after which the feature was dropped. Concern for privacy was built into the culture of the network in its first years rather than added after the fact.
The service layer of Minitel was arranged in a decentralized way: third-party providers connected hardware of their own, Unix machines, mainframes, a Commodore Amiga, ordinary IBM PCs, through the X.25 protocol of the Transpac network. Its competitors of the same generation (Prestel in Britain, BTX in Germany, The Source in the United States) kept all of the content on one central computer.
The future billionaire and founder of Free/Iliad, Xavier Niel, started out precisely with Minitel Rose: he launched an erotic messagerie straight out of school in 1984, and the "3617 Annu" service he created in 1996, the first reverse telephone directory, brought in more than 70 million francs a year and financed his later move into ADSL internet service.
The program survived a change of president and of political camp without stopping: the report that justified it and the decision itself came from the right-wing president Giscard d'Estaing, while the mass distribution of terminals in 1983 was carried through by Mitterrand's socialist government.
Running a chat line on Minitel Rose was almost a profession: providers kept a staff of animateurs, men writing under women's names, plus ghost bots, an artificially inflated count of users supposedly online at once, and deliberate delays in displaying pages so that the customer stayed on the meter longer.

Legend vs. the record

Legend: Minitel was free. That is how it is remembered, as the state simply handing out terminals. The record: what was free was the terminal (it paid for itself in printing savings on directories) and the basic electronic directory (code 3611), while every Kiosque service from 1984 on was charged by connection time straight on the telephone bill, with the revenue split roughly a third to the operator and two thirds to the provider, and it was those payments that made up the network's multi-billion turnover. Verdict: the hardware and the utilitarian minimum were free, and the substantive economics rested on per-minute billing, a model closer to a free reader with paid content than to a free service.
Legend: Minitel delayed the arrival of the internet in France. The formulation is close to a truism in English-language retrospectives. The record confirms half of the thesis with figures: in 1998 only 2% of the French used the internet, against 13% of households in Britain and 60–70% of the population in Northern Europe by 2001, so France was objectively behind. Verdict: the lag is real, but no source offers a controlled comparison of causes; that the cause was the success of Minitel rather than the price of access, the language, or the structure of the provider market remains a thesis of journalism rather than a demonstrated causal link estimate.
Legend: a closed state network. If the network is state-owned and the billing is a monopoly, then all of the content must be under the control of Paris. The record shows the opposite: any provider could lease a 36xx code and connect hardware of its own, from a Unix server to a Commodore Amiga, through a standard protocol, whereas the British and German equivalents kept content centralized on a single machine belonging to the postal authority. Verdict: only the network and payment layers were state-run; the service layer was an open platform for thousands of providers, closer to an app store than to the monolith of its competitors.
Legend: the pink chat lines produced most of Minitel's revenue. Cultural memory (the Houellebecq novel, the series "3615 Monique") fixes exactly that image. The record: messageries roses and games together produced 780 million francs, 35% of Teletel revenue including tax, of which erotica accounted for 480 million francs; the year of these figures is not named in the primary source unverified. The remaining ~65% of revenue came from the directory, banks, tickets, the press, and professional data. Verdict: the most visible and most profitable vertical, the engine of the network's recognition, but not the source of most of the money estimate.

The first growth lever

The lever was not paid content but a free and boring utility. The electronic directory (Annuaire électronique, code 3611) was tested in Saint-Malo and Ille-et-Vilaine in 1980–1981 as a replacement for the paper directory, whose printing and updating was in itself a noticeable expense for the state. The economics were simple: giving away a terminal along with the telephone line cost less than printing and mailing a paper volume every year. Savings on paper reached 500 million francs a year, the whole program cost 8 billion francs in terminals over the first eight years (with an average terminal lifetime of about 8 years as well), and it brought in 3.5 billion francs of turnover in telematic services. Already in the first extended experiment, Vélizy in the autumn of 1980, 2,500 households had 190 services from ~200 providers available to them: from day one the network was designed so that anything at all could be bolted onto the boring directory.

That box, already installed in every home, made the next step instantaneous. When Kiosque launched in 1984, an entrepreneur needed neither hardware of his own nor an explanation to the customer of how to pay; leasing a 36xx code was enough, and the telephone company took the billing on itself in exchange for its share. The millions of terminals already handed out became a ready market for any new service, from SNCF and the banks to erotic chat lines.

Parallels today (projects from the catalog)

Ka Ching (ka-ching-shopify-apps): a portfolio of Shopify apps at $4.5M ARR lives by the logic of Kiosque in pure form. The developer needs no billing and no audience of his own, the store installs the app in a couple of clicks from the App Store, and the platform runs the payment and keeps its share. The difference in eras: in 1984 Kiosque connected providers by hand, by fax and telephone line, while Shopify does the same thing programmatically and instantly. view this project's dossier →
Zigpoll (zigpoll): about a third of sign-ups come straight from the Shopify App Store, a marketplace that brings in the customer and takes a percentage for ready-made distribution and billing, so a solo founder does not have to talk a merchant into entering a separate card for one more service. This is the same barrier to entry that Kiosque removed in 1984, when a provider no longer had to invoice on its own. view this project's dossier →

What a builder can take from this in 2026

Subsidize the boring utility that puts the device in people's hands, and make your money on the layer that arrives after it. Minitel was given away free in order to replace the paper directory, and that installed base in millions of homes made Kiosque instantly valuable to any new service in 1984. A cheap entry point that solves a boring universal problem is still a cheap way to seed a distribution base for whatever you build next.
Own the payment relationship rather than the content. France Télécom did not have to invent a hit service of its own; it took its share (about a third) of everything somebody else built, on exactly the model of a modern app store or payment processor. If you already hold the customer's payment instrument, you can let a thousand others build the product and carry its risk.
Loud content close to the line will finance the infrastructure, so plan for it in advance. Messageries roses were a minority of the revenue but a majority of the recognition, and the same entrepreneurs (Niel, Maublanc) reinvested that money into respectable businesses, Free and Aquarelle. The most profitable early vertical does not have to be the one you will want on your home page ten years later.
A network closed from the outside can be an open platform on the inside. The advantage of Minitel was that third-party providers connected hardware of their own through a standard protocol instead of living on a single state mainframe, as with the British Prestel or the German BTX. That is the difference between an app-store model and a monolith, and it explains why Minitel outlived its European contemporaries by a decade.
A profitable product does not have to lose to a new paradigm in order to be shut down. In February 2009, 15 years after the birth of the web, Minitel was still recording 10 million connections a month; the shutdown in 2012 was an owner's decision rather than a death from irrelevance. Do not write a product off just because a newer paradigm has appeared. Check the usage numbers first.

Discrepancies and what we could not verify

The Kiosque revenue split. Four versions: 15/85, 40/60, a third and two thirds (Stoner 1988 plus a 1998 calculation), and 50/50 (Polytechnique, 1997). The best supported is a third and two thirds, which matches the Kiosque tariff at launch in February 1984 (60 francs an hour: 40 to the provider, 20 to France Télécom). The likely reason for the divergence is different tariff codes (3614–3617) and different years.
The cost of the program. The figure of 7 billion euros is repeated in popular retrospectives with no breakdown and is a likely franc-to-euro conversion error. The consistent figure, 8 billion francs for terminals (3.5 billion of service turnover, 500 million a year saved on paper), is independently confirmed by the French and English Wikipedia and is the one used here.
The date of the Saint-Malo experiment: July 15 vs. "consultation launched on July 29" vs. simply "July 1980" in a third source, probably different steps of one process; the timeline gives "1980" without false precision.
Who founded Gretel, and the voice of a Kiosque service owner: resolved as Michel Landaret, head of the computing service at Dernières Nouvelles d'Alsace in Strasbourg (the name that had been circulating was confirmed nowhere). In his own account (Slate.fr, 2011), the first public chat was born by accident, when a ten-year-old boy cracked the password to the moderators' messenger (1981) and passed the method around to his friends, after which DNA opened the feature to everyone. That is the sought-after account from the owner of a service, only of the earlier Gretel rather than of Kiosque; for Maublanc and Niel there are still only retellings.
The year of the figure "780 million francs / 35% of Teletel revenue." The sum and the share are confirmed by a primary source, a question from a deputy of the National Assembly, but the year is not named in the document; the 1987 given in secondary sources is not directly confirmed unverified.
The full text of the Mailland and Driscoll book (MIT Press, 2017) is behind a paywall, abstracts only.

Sources (primary first)

Secondary (facts, figures, and quotations used after checking against several sources):