The coalescence of data-based culture

There’s been a debate in France last year about a law that prevents online book shops (i.e. mostly Amazon) from discounting their prices too much. This is clearly intended to ensure that traditional brick & mortar bookshops don’t disappear because they can’t withstand that kind of competition. Of course every libertarian railed against this meddling of the Government against the will of the Free Market. I’d like to explain here why I think this law is a good thing, and to reframe that in a more general problem : audience data has become the only criteria which drives culture, and this is turning it into a sterile endless rehash of the existent.

A couple of decades ago, if you wanted to buy a music record, you’d go to a record store. At that time, there still were quite a few of those. Since then, they’ve been killed by very large record stores (FNAC and Virgin in France, Tower Records in the US) and big malls like Carrefour or Walmart who sell only a few very popular records at a very low price. And now those large record stores are in turn being killed by Amazon and online music services like Spotify. What the record market lost with those record shops is a very good curation layer that was constantly listening to new stuff and would promote it to their customers whenever they liked it. That’s how many artists emerged, against the market trends that were current at their time. Contrast to what we have now, where the music landscape has a high turn-over of artists who don’t have time to establish themselves before vanishing from sight, and yet the musical landscape remains fairly stable, as most artists sound like many other.

This is the simple consequence of the fact that all record companies now only follow the sales or listening data they get from the market to determine their next move. And the recent arrival of big data providers like Amazon or Shazam make this only more pronounced. The data point to a short set of music types that people tend to prefer, so that’s what music majors provide, rarely risking to deviate from that guideline. In turn, people are mostly exposed to only this same set of types, which of course increases their popularity, and you’ve got your feedback loop. It’s the same as what a “top 10” listing displayed on a front page will do.

So the result is a big cultural precipitate, a race toward what the data says people with an ever-shrinking musical landscape will like, with no one to stir the pot, keep things fluid and breaking lumps. From Amazon’s algorithmic recommendations to what the music industry itself promotes, everything tends to put people in a cultural bubble from which they stand very little chance to pierce through. But since humans are easily bored creatures, they tend to lose their interest, and there you have one of the reasons why music is losing its cultural importance.

This is partially counter-balanced by the availability of a huge amount of music on the Net, but there’s no one to guide you through it. You can follow recommendations from friends and families, but there’s no one with an actual in-depth knowledge of musical culture who’s able to tell you “try this, it’s completely different from what you listen to, but it’s important”.

Likewise, this now applies to TV as well, as cable, satellite and on-demand video services can give tons of very precise consumer data.

You may have seen the movie “Mondovino”, a very interesting documentary about the wine-making industry. Wine is probably among the most territorially-based product, you literally drink what a very specific piece of land has produced, each with its own traits due to the soil, the vine, how the grapes are processed, and overall the technique of the wine maker. Yet, this variety of tastes is disappearing, because every wine-maker is trying to get a high mark in the most prominent wine guide, which means every one is trying to achieve the same “most popular” taste.

Finally, I wonder if this phenomenon also applies to politics : following market data, media becomes polarized, in turn people’s mindset become polarized as well, and then so do politics in general because politicians essentially cater to a specific audience which they study through polls.

To conclude, this is one (yet another) case where blindly following the Market is profitable in the short term but destructive in the long term. You need to maintain a layer of people whose job is to be attuned to a given branch of culture and who will have some amount of influence to promote really new stuff. Algorithms won’t replace that anytime soon. If one does, it’s likely to pass the Turing Test as a side feature.