At 9:04
Many mornings, at 9:04, Yves Bigot would gather his things. The previous night's ratings had been bad and he was convinced they were going to fire him. It didn't happen just once. «Si vous saviez le nombre de matins…»: if you knew how many mornings.
He had spent decades around music and the media: writing about records, programming songs, working at record labels, in radio and in television. He had learned to recognize an artist, to defend a show, to make decisions aimed at very different audiences. Even so, some mornings all that experience ended up facing a curve.
The figures reached the research department. They examined the show's ratings, its competitors' ratings, the behavior of the different age groups, of men and women. Then the interpretations began.
And the interpretations had owners.
The advertising régie (the people who sold the channel's advertising) had an opinion, like everyone else. But it didn't give it to him. It gave it to his superior, «pour qu'il en tire les conséquences», so that he would act accordingly. From there came convictions, trust, the company's internal politics, the press and, in public broadcasting, the political world as well.
«The figures are always interesting and should be part of any reflection», says Bigot. «They become decisive when they are bad!»
Long before a platform could record every song we listen to, how long it takes us to abandon it or what we play next, culture had already learned to see itself in numbers. And those who decided what should reach the public knew that those numbers could end up deciding about them too.
It took Bigot a while to understand to what extent.
J'étais naïf
J'étais naïf: I was naive.
Bigot was not unaware that a record label is a business, nor did he think the people who worked in one were philanthropists. What surprised him on crossing from criticism and music journalism to the other side was the concrete weight of the financial factor and, especially, the artists' own relationship with their income, which often mattered to them as much as their artistic expression. Something that, he admits, is legitimate, and that drove many more decisions than he had imagined.
With programmers something similar happened. For a long time, he says, he had «l'incroyable chance» of being shielded from that pressure. Later he understood the «dictature de l'audience», which obliges you to be effective above all else, on pain of losing your job.
The 9:04s came later, but the seat was already unstable. In 1995 he spoke in Libération about how hard it was to build careers when a project might need five or six years and the people running it lived on a siège éjectable, an ejector seat. He also described an audience «terriblement sollicité», besieged on all sides, with an enormous capacity for forgetting.
At the labels he had also learned that recognizing an artist's worth and getting the public to notice them were different skills. He had thought that good artists backed by good records would end up finding buyers; perhaps not immediately, but in time, feeding a catalog that would still be worth something years later.
Reality had a lot less patience. TV advertising, incessant deals and the dictatorship of the Top 50. From that experience he keeps an axiom that he himself calls terrible: «What works can't really be bad, and what doesn't work can't really be good».
«What works can't really be bad, and what doesn't work can't really be good».
He knew the mechanism well. Between 1985 and 1988 he had replaced Marc Toesca on Canal+'s Top 50. Years later he could talk about that system from the experience of someone who had been part of it.
There was another fragility that no chart reflected. Companies were made of people, and decisions could leave with them. Bigot mentions Marc Morgan, a Belgian singer, «très prometteur», whose first record had come out on Fnac Music in 1993. According to Bigot, when he left Mercury, the company abandoned him completely. To describe it he uses an American expression: spoils system.
Asked what the music critic he once was would have reproached the executive and programmer he became, the answer leaves little room for nuance: «d'évidence», for not having taken enough risks, for having bowed too much to the market and to the Audimat.
In 2005 he had put that transformation another way. Asked whether he would schedule rock in prime time on France 2, he replied that he tried to be «professionnel, pas militant», professional, not militant. If he put a rock show on in prime time, he explained, it could make it into the Guinness book of the lowest ratings and ensure that for ten years no channel head would want to hear about rock again. Numbers had a memory too.
The critic who had spent years recognizing what deserved to be listened to ended up occupying places where recognizing it was no longer enough. It had to be sold, scheduled, justified, and one had to decide how long it could be granted before accepting that it didn't work.
How does it feel?
When we ask him about his own ear, Bigot begins with a phrase in English. «La seule chose qui compte pour moi, c'est "How does it feel?"» The only thing that matters to me is «How does it feel?»: what does it feel like?
The question is Dylan's, from 1965.
Bigot knows that this ear isn't neutral either. He talks about his «oreille de boomer» and admits that, «sûrement», it influences what he feels. For a long time he expected rock to keep a dominant position and, above all, to remain a force capable of changing the world.
When he tries to explain at what point he understood that rap and EDM had taken over, Live Aid and the Nelson Mandela concerts come up. Those events, he says, prolonged the illusion that rock would keep that role.
What he was losing was also a political expectation placed on music. In 2016, in an interview with L'Orient-Le Jour, he was already lamenting that «la dimension politique» had been lost.
That subjectivity can make it harder to recognize the new. It also allows one to insist when the data still advise otherwise.
In October 1999, Bigot ran the variety, game shows and entertainment unit at France 2. Among the shows he launched was Un gars, une fille, a daily fiction starring Alexandra Lamy and Jean Dujardin. It was, he says, a «complètement novateur et casse-gueule» format. Two months later, Bigot recalls, practically everyone wanted to pull the show. The problem, according to him, was «les risques qu'il faisait prendre à l'information»: it aired right before the eight o'clock news.
There was one exception. Michèle Cotta, director general of France 2, did not join that pressure.
He places the change around the following Easter. Un gars, une fille ended up staying on the air until October 2003 and became one of the great hits of that period at France 2. It took at least two people: someone willing to defend the show and someone above him willing to give it time.
Speaking about the figures, Bigot had written: «Après, tout est question de convictions, et de confiance». After that, it's all a matter of convictions and trust. Trust was an infrastructure. It made waiting possible.
«After that, it's all a matter of convictions and trust».
There is also an unavoidable bias in the stories that survive: we remember Un gars, une fille because it worked. The shows that someone defended against the figures and that ended up failing rarely become examples of intuition. The numbers described well what was happening at that moment. They couldn't say how much time to give to the possibility that something else might happen.
That is the space Bigot will reserve, later on, for the programmer.
The Machine Without Affect
We put a hypothesis to him: if tomorrow there were a system capable of recommending music and getting it right almost every time, what would a human programmer still be doing? Bigot immediately reframes the question: «You raise the essential question: is AI superior to the human being?».
In that scenario, he says, the machine's superiority lies in its «manque d'affect, its lack of affect»: it will not favor a friend or a head of press and, in all likelihood, will not lend itself to corruption either. It is a considerable admission from someone who has spent a good part of his life inside the institutions that selected, promoted and programmed culture.
The machine, he says, has no emotions and, «pense-t-on», no interests either. Or so it is thought.
On legitimacy, Bigot has no doubt: «La seule légitimité appartient au public». Companies and media have it only over their own airwaves: «elles sont chez elles, elles font ce qu'elles veulent». The State «doit — devrait —» guarantee equal opportunity and diversity.
Faced with a machine that got it right almost every time, Bigot reserves for the programmer what prediction can hardly guarantee: intuition, vision and the possibility of introducing something new, different or disruptive.
Affection can lead one to favor a friend or to listen too closely to a head of press; it can also sustain a conviction when there is not yet any data able to back it. Affection contaminates human judgment and makes human judgment possible.
In 1995, Bigot compared investing in French song to a game of poker. Whoever didn't invest in it, he said, was like a player who didn't bet. Three decades later we can build systems capable of calculating probabilities with a precision that player didn't have. It remains to be seen whether betting consists only of doing that calculation better.
In 2024, still at the head of TV5MONDE, Bigot saw artificial intelligence as a tool for specific problems, such as subtitling and dubbing. He is not a man who needs to imagine it from the outside.
But when we ask him which part of selecting, contextualizing or betting on a work he would be willing to delegate to it, he answers:
«Je ne lui confierais rien du tout! Ou alors ma place. Sinon, c'est que je ne suis pas la bonne personne.»
I wouldn't entrust it with absolutely anything. Or, if anything, my job. Otherwise, I'm not the right person.
«I wouldn't entrust it with absolutely anything. Or, if anything, my job.»
Three Songs
Before finishing we asked him for something different: three records that hadn't come up during the conversation and that said something about him. Not his favorites or the best ones, but three titles that portrayed him today. And a why.
He preferred to answer with three songs: «Not Dark Yet» by Bob Dylan, «Second Song» by Neil Young and «Days» by The Kinks.
«Not Dark Yet» came out in 1997, on Time Out of Mind. Dylan was 55 when he recorded it in Miami, with Daniel Lanois producing, an organ and two drummers. It lasts six and a half minutes and addresses no one: a man speaks before an ending that is approaching but has not yet arrived.
«Second Song», by Neil Young & the Chrome Hearts, was released just a few weeks before Bigot answered the questionnaire. Young is eighty. It is the title track of an almost entirely acoustic album, lasts more than eleven minutes and revolves around memory, loss and the need to find someone his own age to share memories with.
And then, «Days». Ray Davies was twenty-four when it was released in 1968. A farewell turned into thanks to someone for the time shared. Davies would link it, many years later, to his sister Rosie's departure for Australia and would recall that he came to think it might be the Kinks' last single: «Pop musicians aren't made to last forever».
A voice of 55, another of 80, another of 24.
We had asked him why. He answered with three titles.