My algorithmic purge

In the last few years, I got divorced and moved town and... wait, come back! This isn't about that. I'm going somewhere here. It hit a point where Netflix cut me off for not being in the same household as my kids and so I had to get a new Netflix account. It was a clean slate. An algorithmic purge. And wow, that was an eye-opener. Because suddenly I was getting served up all kinds of things I never knew existed. Stuff that, on my old account, the algorithm had decided wasn't for me based on my previous viewing. And, sure, with a good chunk of it the algorithm was right.

But not all of it.

I found some good stuff that, otherwise, I'd never have seen.

And I suddenly wondered just how limiting the algorithm might be.

In recent years, I have found it trickier to predict which films will hit and which won't. I feel I used to be better at that and I do remember wondering if I'm just losing touch with the zeitgeist. It was when an industry colleague mentioned the sheer amount of advertising for that Lord of the Rings animated movie that it really hit me. I hadn't seen a single piece of advertising for that movie and I'm still not sure anything existed outside of LA. But how could I know? We now live in little algorithmic bubbles. I loved Tron Legacy back in the day so it's probably no surprise that my social media feeds were loaded with Tron Ares ads when that film was close to release. To me, that movie was on my little doom handheld every time I opened it up. I was bombarded by advertising. Others saw nothing. Not a thing. The movie did not do well. Other movies hit huge and I think, I didn't even know that movie existed.

It's not so much that people (me, anyone) are losing touch with the zeitgeist. It's that collective culture doesn't exist the way it once did. When I was a kid, I was never into golf. I'm still not. But I had heard of Arnold Palmer. I knew Hulk Hogan (I know he's not a golfer, I've moved on, keep up here). Anyone remotely famous in any field was cross-pollinated across mass media. Now, I couldn't name a single golfer. I can name wrestlers but I'm guessing many couldn't. We're only served up the things we've already shown an interest in.

And it works. It works really well. In some ways.

Targeted advertising is great at efficiency. It is known for good conversion rates because you're essentially preaching to the choir. I'm guessing it is a much lower cost per acquisition in a lot of cases (marketing people can tell me if I'm right or wrong there) and, of course, feedback is likely much faster. You're selling to people already interested. Capturing existing demand. The algorithms can optimise for who is most likely to convert now.

But maybe not for who might become a customer in the future.

It is harvesting demand, not creating demand or even exploring demand. The algorithmic bubble is, in ways, actually keeping us blind. It is preventing the accidental discovery. Creating a wider cultural presence and really expanding the market, can that be done in those little algorithmic bubbles? If you're only pushing things to people who would have bought it or watched it anyway, reaching the same group again and again, doesn't it risk a misleading sense of cause and effect? I find this on Meta platforms a lot - ads for things I'm buying anyway. Somewhere there is a data analyst thinking, this worked. We served up an ad, he bought the product. But the ad is being served up to many people who had made their buying decision already. So it's possible, even likely, that the more targeted those algorithms get, the more you have just a broken feedback loop that keeps niches narrower rather than growing.

I think we may have a little problem there.

What happened when I had a total Netflix refresh was more akin to the old way of marketing. It didn't know who I was. Not exactly. It couldn't know what bubble to shove me into. It couldn't just serve up only what I would have looked for anyway. And so it scattered petals into the wind. I saw some unexpected choices. My eyes were opened. Serendipity happened. And for a brief moment on Netflix, one or two shows found a new audience in me.

And that's the thing about data. It can be so useful and powerful in connecting interests, in matching people to trends. It's important. But finding new audiences? Completely new ideas? Wild card films and television that could hit big? We might need to scatter more petals into the wind.