Note: Netflix is the example used as it’s the most well-known streaming platform, but most of what is set out below can be directed towards Amazon Prime, Disney+, Hulu, or any other streaming service you use.
How often have you switched on Netflix looking for something to watch? And how often did you end up scrolling endlessly before turning on a show you’ve already seen, just because you couldn’t choose? Don’t worry, you’re not the only one. A recent Nielsen report, one of the primary sources for audience measurement information in the US, found that most people take over 7 minutes to decide what to watch, jumping to over 9 minutes for people between the ages of 18-34.
This isn’t anything new. The term “Overchoice” was first coined way back in 1970 by Alvin Toffler in his book ‘Future Shock’. Overchoice, or choice overload, happens when consumers are faced with an overwhelming number of choices. The vast amount of outcomes overload our senses and we fear of making the wrong choice.
Since Toffler first published his book, we have become more and more inundated with choice. Unless we already have a show or movie in mind, we find a hard time picking what to watch. But why do we have a hard time deciding what to put on? And is there any way to fix this?
How does Netflix deal with overchoice?
You’ll be happy to hear that Netflix isn’t blind to this issue. It currently has complex machine learning systems that try to make choosing easier for you. They make use of the massive amounts of information Netflix has collected on users to create recommendations based on what movies and shows you watch. But they’re not actually recommending the show to you. Rather, their system is estimating the probability that you’ll like a show. If you watch a show like Arrested Development, and 90% of people who also watched Arrested Development went on to watch Schitt’s Creek, Netflix will pick up on this trend and recommend Schitt’s Creek to you.
This issue is so important that Netflix has dedicated research just for personalisation. It is essential to Netflix that each user feels sated in their desire to watch programs. But there is an issue, it’s an issue that Netflix addresses themselves in the third paragraph on their research page.
Of course, dealing with human tastes and preferences is an extremely challenging problem. In many cases a member may come to our site not knowing exactly what they’re in the mood for. Each person is nuanced in what brings them joy and how that varies based on what context they are in. Plus, people change over time, so we need to help anticipate or enable those changes. We also need to supplement our algorithmic approaches with insights we gain from consumer research and curated metadata to incorporate more of a human element in our systems.
Netflix invests a lot into machine and computing based solutions for its personalisation software. It attempts to predict what you want to watch before you want to watch it. But Netflix can go a lot further than giving you more personalised options. In fact, Netflix’s current attempts at personalisation might even be making it harder for you to choose! As one study puts it:
Empirical evidence suggests that consumers are less likely to prefer large assortments over smaller ones if they assume that the options in both assortments are mostly attractive and of high quality.
If we were given 100 bad shows and 1 good show, the solution would be easy. You pick the good show. The problem of overchoice comes in when we’re shown 100 good shows and asked to pick just one. In Netflix’s attempt to create a solution to the difficulty of choosing a show, they’ve paradoxically made it harder than before.
So where does Netflix go wrong? How can it change its service with the goal of decreasing overchoice frustration for its users? Let’s dive a little bit more into Overchoice.
How does overchoice happen?
We can break overchoice down into two sections : how our content is arranged, and our interaction with the content. First, let’s look at how Netflix arranges content.
Netflix puts shows on a vast main page, with near endless scrolling to reach the bottom of their offered lists (which all have even more options we need to scroll to see). It can be incredibly intimidating. Outside of Netflix original content, all programming is given the same size and importance. It’s difficult to decipher between options when they are all at the same level in the hierarchy.
However, Netflix categorizes its content really well. It breaks them down by genre (sometimes incredibly specific genres), gives you lists of similar shows to previous ones you watched, and also offers a variety of individual lists, such as popular shows, recommended shows, and shows to rewatch.
So Netflix does a reasonably good job at categorizing its content, but suffers when it comes to the arranging it all.
More important than just the arrangement of content is how we interact with it. We can look at this in two parts : when we’re trying to make a choice, and the aftermath of our choice.
When we’re trying to make a decision, we often set out with a goal in mind. We might be fully conscious of this goal, like when we set out to watch a specific show, or it might be something more vague, like wanting to have a show on for background noise. When we’re looking at helping people overcome choice overload, we need to be aware of their goals.
Now that we have our goals, we need to look at the information that’s given to us. Netflix gives us a healthy amount of info on each show : a show summary, top actors, director, and how well it matches with our Netflix viewing profile. Information that is missing that might be helpful for us is what other people think of the show. Maybe is matches our viewing habits, but is it any good? Netflix gives no information on quality on its platform.
Finally, we have to look at how long we spend putting off a decision. As mentioned right at the beginning, most people spend between 7 and 9 minutes trying to decide on something to watch. The length of time we spend putting off a decision is a great indicator of our difficulty choosing an option, and therefore is a great bellwether for when we’re improving or worsening overchoice.
After we make our decision, we look at how confident we are with it. Do we regret our choice, or are we satisfied with it? This influences if we decide to continue watching a show/film, or if we turn it off after the first five minutes.
We also look at how we justify our choice. Often when we are faced with overchoice, we will pick the option that’s most easily justifiable (a.k.a the sitcom you’ve rewatched dozens of times).
So what does this all mean?
Now that we understand the way overchoice works, we can start to examine Netflix’s system and look at a few ways we can improve it.
Netflix could rearrange its content so that its recommendations to you give off more importance. This could be done a number of ways, such as making them a larger size (what Netflix already does with its originals) or by putting them on their own page and moving the long page of lists to another heading, such as “discover”.
They could make new categories that better reflect the user’s goals when they log in. For example, they could feature movies that are a certain length, or categorize by emotion or intensity. Personally, I’d like to see a category with shows I don’t have to pay full attention to!
A big issue is figuring out if the show we watch will be good or not. A lot of the time, we work on recommendations – not from Netflix, but from our friends and general public consensus. Netflix could introduce ratings from an external source (although this could lead to some content never being watched), or they could take it a step further and introduce an entire review ecosystem, where people can follow their friends or notable reviewers and gain recommendations from them.
These are only a few examples, and there is a lot more that could be done. Think about it for yourself, and let me know your ideas.
Overall, I believe that Netflix concentrating solely on their machine learning system will lead to future problems. There are too many issues that aren’t addressed with this solution. There are a lot of other features and tweaks that could be used to reduce the issue of overchoice.
Or maybe the issue of overchoice will never go away, and we just need to get used to those few minutes we spend trying to find the right content.
Bibiography
Can There Ever Be Too Many Options? A Meta-Analytic Review of Choice Overload
Choice overload: A conceptual review and meta‐analysis
How Netflix Uses Analytics
Netflix Research – Recommendations
More Is More: Why the Paradox of Choice Might Be a Myth
Futureshock – Alvin Toffler