Anyone who has flipped through their favourite streaming services after a long day at work knows the frustration of the endless scroll. The question is often not: "What is my taste?", but rather something subtly more important: "What am I in the mood for, exactly?"
Enter Netflix: which is working on solving the "mood" problem.
The company confirmed to The National the project is under way. It began by trying to change what Netflix recommends based on the time of day.
Now, developers are working on adding options that would let users tell the platform how they’re feeling, which would change what content it displays.
“The thing is, we all do have very diverse tastes,” Patrick Flemming, senior director, product management, says.
He adds: “It could be film night, it could be a series night, it could be a game night. The question for us becomes, how do we get even better at discerning what you’re in the mood for right now, so that we can adapt those recommendations accordingly?”
Flemming says he and the Netflix product team have made figuring out how to recommend based on mood one of its top priorities for the year. However, due to the complexity of the issue, the answer may not be found in the immediate future.
“This journey is probably going to be a couple of years,” he adds. “But this is a big focus for this year that we’re trying to get a better sense of right this moment.”
Netflix is not the only consumer-facing tech company seemingly trying to solve the issue. This month, YouTube reportedly began rolling out pop-ups that prompt users to select a colour from red, blue or green, which it will seemingly use to train its recommendation algorithm. Users' choices reflect the suggested videos on their feeds.
The problem is, how is a platform such as Netflix supposed to be able to guess how we’re feeling? One idea it tried in the past, Flemming adds and implies they are already using, is changing what it recommends based on the time of day.
“To be direct, [time of day] is something we’ve either tried to incorporate before or may already incorporate in terms of the way we would consider what to recommend you,” Flemming who leads the company's product team responsible for member experience, adds.
Part of the reason that this gets difficult is the same reason that Netflix’s algorithm is already so effective – the company tends to get quite granular when judging success and failure. For instance, the streamer knows when you scroll past something, and views every skip as a failure of sorts.
It shows you a title and an image in a little rectangle and if you do not select it, that means it was not enticing enough to get you to click, in the company's view. That’s why those images may change – as nearly every title has different images that the platform A/B tests to see what is most effective at visually enticing its users to choose it. Everything is constantly being optimised.
The frustration for Netflix’s product team comes in the unknowns.
“When you open Netflix, we’re going to take our best shot,” Flemming adds. "But that’s with no context for who’s sitting with you or what mood you might be in.
"We can’t sit there on the couch next to you, but we can offer you good ways to tell us this, based on what you’re browsing, or what you’re searching for."
His team is now looking at how to allow users to tell the platform their mood, which will allow it to adjust accordingly.
“There might be a way that we would offer you a little bit more explicit input … it could be a ‘family movie night’, or a ‘looking for laughs’ night,” he adds.
“There’s nothing to specifically reveal on the direction we’re going to go, but I can say it’s a keen area of interest for us to better on that front. There’s no news to break on timelines, but we’re working on it.”
