We Asked A Data Analysis Expert Why Streaming Services Recommend Us The Weirdest Gear

Sometimes, you just want to watch a movie that you KNOW you’re going to actually like. Browsing through catalogue upon catalogue of options can be overwhelming and if you’re anything like me, chances are you’ll end up picking something you’ve already seen just because it’s easier. I’m lazy, okay?

But then we have the recommended tabs. The options that streaming services think are best suited to your viewing habits. And while sometimes they can get it absolutely right, other times… Not so much.

So we spoke to an expert on data, to find out exactly why streaming services sometimes throw a totally out-of-the-box bone-tingling horror flick your way when all you want is some romantic and lame Christmas movies.

my face, watching any soppy romance

Dr Gentry White is a Senior Lecturer in the School of Mathematical Sciences at Queensland University of Technology (QUT) and he is the kind of person that can help explain what’s going on here. He describes his work as:

“As a statistician I specialise in data analysis and inference, which is the analysis of data to both reveal underlying patterns in noisy data, and to make statistical inference about questions about the underlying processes that produced the data.”

Sounds complicated, right? Well, it’s actually pretty straightforward. Basically, what he does is wade through complex and noisy data to find the little nuggets of information that make patterns, and question the numbers behind how the data was produced.

Which means he was the perfect person for me to interrupt his day and pester about why I, a certified chicken, keep getting suggested creepy horror and action movies.

my face when I spot a ghoul while looking for soppy romance 

My understanding of it (though limited in nature) was that it looks at the genres and types of movies you watch, and then makes an assumption based on other movies and shows with those same ‘tags’, so to speak. And, to my surprise, I wasn’t completely off base.

“That is a pretty accurate description,” said Dr White (YESSSS).

“But they also tend to use additional information, like what other movies or TV shows people watched together.  For example, if a lot of people who watched Aliens also watched A Walk to Remember. If you watched Alien, you might be recommended other sci-fi movies, but you might also be recommended A Walk to Remember.”

Long story short? While you’re sitting there feeling special that your feed of recommended was curated just for you, that may not always be the case.

“It’s like when you go to Amazon and get [the alert] ‘people who bought this also purchase that…’,” says Dr White.

it’s fine, I’m fine.

And it’s the same when you consider the types of random ads that pop up on your social media feeds. Earlier this year we wrote about how the online retailer Wish markets a whole bunch of bizarre stuff, and it got me questioning whether it was based on the data that people had given, or more of a data grouping like with streaming.

According to Dr White, “Companies like Google specialise in tracking what you search for and then directing targeted ads at you when you use their services. As an example, if I am shopping for something online, then invariably ads for those products or similar ones will show up.”

“Social media companies are essentially like advertising agencies, they make money by selling ad space, and part of their appeal is that they can target ads to specific audiences based on their interests, like their internet search history, to increase the likelihood that people will buy products or services.”

He did also agree, however, that there’s a chance that Wish specifically chooses to advertise the weirdest products as a marketing ploy to get people talking. If so, maybe my colleagues aren’t as bizarre in their online shopping choices as we thought.

Jury’s out.

If you find all of the data and analysis as fascinating as I do, you can study it yourself at QUT. You’d get to study real-world data sets and maybe even let me know whether or not the stuff I Google really does impact on why Wish kept sending me random ads.

Because the bottom line is, in writing this, I was served an ad for leggings patterned with Nicolas Cage’s face.

And maybe it’s the data getting it right, but I kind of want to buy them.

More Stuff From PEDESTRIAN.TV