Let’s fix the “Netflix Problem” the right way.
The streaming market is growing. A research institute Fortune Business Insights projects the global OTT market will grow at a CAGR of near 20% to exceed USD 1,690 billion in 2029.
However, the market environment isn’t all peaches. The competition is heating up, driving up the cost of retaining paid viewers, all the while most mainstream OTTs are faced with the “Netflix Problem.”
What’s the Netflix Problem?
Original shows helped OTTs attract new subscribers. Today, the abundance of content itself is deteriorating the user experience.
There are simply too many films and shows on the platforms and viewers are finding it difficult to navigate their way to find the right ones to watch.
The likes of HBO Max are trying to solve this by narrowing down their target users and taking down less-watched shows. Others like Netflix hire human taggers, to categorize each content in an attempt to help viewers find shows they’re likely to enjoy.
Neither really solves the problem. HBO Max can’t keep taking down old shows as it brings on more shows to stay competitive. Netflix, on the other hand, recently proved the problem with its tagging method — it relies too much on the individual tagger’s points of view.
Could they use the Spotify Model?
Some people suggest that Netflix should take on the Spotify model — using API to compare the user’s streaming history and recommend accordingly.
This could be an option, but one that creates the “cold start” issue. It takes time for the algorithm to learn what each user prefers, so it’s useless until the algorithm “warms up” enough to be able to recommend the right content.
Also, we don’t watch movies and TV series the way we listen to music.
Most Spotify users (or any other music streaming service, for that matter) wouldn’t mind having a few ‘irrelevant songs’ in their playlists. Why? Because you can simply skip to the next one on the list, or do something else while you’re not listening to that song.
Either way, it doesn’t really impact the user experience negatively. Most of the time, we play music in the background while we do other stuff anyway.
Streaming movie is different
It takes as short as 30 minutes to as long as 200 minutes to finish watching a show. Also when you watch a TV show, you prepare yourself to be immersed in the experience.
So it goes beyond ‘disappointing’ to start the play button, only to find out that it’s not what you wanted to watch.
Then you have to decide if you’re going to continue watching it or move on to another one not knowing if you’re gonna like it.
After a while, you find yourself just browsing through the catalog of movies and TV shows for hours and hours… until you finally decide there’s really nothing to watch.
This is a matter of survival and death for many OTTs
OTTs like Netflix and HBO Max spend billions of $s each year to produce original shows to attract and retain subscribers.
In contrast, not everyone pays to watch.
Netflix estimates that more than 100 million households get access to its service using someone else’s password. So it’s looking for ways to stop this — like offering the ‘extra homes’ option at additional cost.
Will this solve the problem? Not if it’s not about the cost.
Subscribers ditch their own paid account and opt to share an account with their family and friends because they don’t find the value in having their own account. They then move on to another OTT, which they share with their friends, and so on.
So really, OTTs should come up with a unique value proposition in having their own paid account to solve the “Netflix problem” and encourage paid subscriptions.
Personalization is the key
Let me tell you about the case of KakaoPage.
KakaoPage is Korea’s leading web comics and novels platform that archives over 70,000 contents — which is the largest in the market.
In 2019, KakaoPage was already leading the market.
However, their concern was that their users — whom they managed to attract by onboarding over a thousand writers and artists — were reading only a handful of best-selling content.
They were experiencing the “Netflix Problem” as hundreds of new content flooded into its platform, making it difficult for the users to discover both existing and new content they’re likely to enjoy.
To solve this problem, KakaoPage partnered with Mycelebs to create an AI-powered personalized recommendation system called “Keytalk”.
Keytalk introduced a powerful new way for KakaoPage to present its entire 70,000 contents to the right users — delivering a 200% growth in its daily transaction value.
What Keytalk did for KakaoPage users is simple. It processed what people were talking about each content, and turned it into an index that could be searched and ranked.
For example, you can see from the web novel you just read it has Keytalk tags like “exciting,” “makes you feel good,” “beautiful artwork,” and “plot with a twist.” You then click on the Keytalk tag you like to see what other novels rank high in similar contexts.
This enables a different level of personalization — one which the user can interact with, explore and build on without having to read the content itself.
Also, there are hundreds of Keytalk tags on any given content, so you’re never limited to a choice between A and B. You can get the feel of the novel (or comics) just by looking at what kind of Keytalks are tagged.
Having a means to discover new content (+engage and contribute to the tagging in real-time), encouraged KakaoPage users to spend more time on the app and read more — and eventually buy more.
That’s how their daily transaction value doubled in a matter of a few months.
The power of context-aware search and recommendation
The same strategy can be applied to most OTTs like HBO Max, Netflix, and Disney+.
This is a scalable solution because Keytalk is AI-powered and fully automated.
No more manual tagging. Keytalk can learn from user reviews and other data sources to update its own data, so even the oldest of the oldest classic movies will always have new, relevant Keytalk tags attached to them.
Keytalk’s curation is based on the context and is led by the users themselves. There’s never any reliance on personal data or user logs, which means there’s no cold start issue either.
And that’s important in this day and age when everyone’s fed up with their personal information being tossed around for the benefit of corporations.
This isn’t where it ends, because Keytalk can even help with the production budget.
Knowing what your users are saying about a particular style of story(plot), character, screenplay, actors, directors, etc., gives you the means to deploy your production budget better. You can make an informed decision based on facts, not guesswork.
Again, there’s no exploitation of personal information here because the OTTs are getting access to the trend, not the user log.
This is extremely important to us at Mycelebs because our mission is to help businesses use AI and data in a way that will benefit and empower the users — the people.
Keytalk is all about empowering the users
Keytalk is an advanced search and recommendation tool for businesses, but our mission is to ensure it empowers people.
We’ve shifted from TV to streaming because it’s empowered us to watch what we want when we want.
Now that’s the norm, the next step should be to empower streamers to discover the movies they want to watch, based on what they like. That’s where Keytalk comes in.
To make this happen, OTTs need a scalable and reliable solution. One that’s neither too expensive nor time-consuming to execute — Keytalk.
This is undoubtedly the most ethical and user-friendly way for OTTs to use data and AI and grow.
