Netflix and the Myth of Personalization

Today Slate has an analysis of how Netflix began the process of personalizing marketing, of using "algorithms to micromanage distribution, not production" in particular in the multi-pronged marketing of House of Cards by creating multiple trailers to appeal to particular slices of the Netflix customer pool, based on their "likes."

In the middle of the article, we find this paragraph:

House of Cards thus embodies one of the most seductive myths of the algorithmic age: the ideal of personalization, of bespoke content assembled especially for each one of us. In fact, the content, or at least the costly, aesthetically rich content we care about, like Fincher’s show, is still fairly limited. There is only one House of Cards, but there are as many ways to market the show as there are to target Netflix viewers. This is what information theorist Christian Sandvig calls “corrupt personalization”: the ways that algorithmic culture blurs the lines between our genuine interests and a set of commodities that may or may not be genuinely relevant, such as products “liked” by our friends on Facebook even if they did not knowingly endorse them.

The piece on corrupt personalization is worth the side trip, but it's a bit much to squeeze in here. But let me toss out three context-free quotes that may ring bells.

It’s as if on Facebook, people were using the yellow pages but they thought they were using the white pages. 

In sum this is again a scheme that does not serve your goals, it serves Facebook’s goals at your expense.

Money is used as a proxy for “best” and it does not work. That is, those with the most money to spend can prevail over those with the most useful information. The creation of a salable audience takes priority over your authentic interests.


And I will bring back Greene's Law-- the free market does not foster superior quality; the free market fosters superior marketing.

Personalized learning, whether we're talking about a tailored-for-you learning program on your computer screen or a choose the school you'd like to go to with your voucher, is not about actual personalization. It's about another path for marketing, a way of personalizing the marketing of the product, the edu-commodity that someone is already trying to make money from.

We're being sold (and in many cases are arguing against) an AI that spits out just the digitized worksheet that Student 12-5452 needs to continue studies, but that's not where we're headed. Look, for instance, at the new, improved PSAT that returns both a score and some recommendations. "Looks like you need to log in to Khan Academy's lesson series for calculus." Or "You would really benefit from the AP Calculus course-- talk to your guidance counselor today."

That's the personalized learning dream-- students with vouchers paying for education one course or micro-credential at a time, and each exercise on the "parent" program ends with, "Good job! You should probably sign up for Edubizwang Corp's Intro to Pre-Pre-Calculus next-- just enter your edu-voucher account number." Marketing that can be directed with laser-like precision at each individual consumer. Marketing that can tell the consumer, "Yeah, this-- this is what you really want."

It's not a personalized product, but the personalized marketing will make you think it's just what you want. Netflix is just the beginning.