Too much choice is no choice at all — the search for the right content
I want to start by sharing a brief anecdote about my afternoon tv watching ritual during weekends. Weekends are usually when I explore the content on Netflix and Amazon Prime. About 2 years back when I started subscribing for OTT (over-the-top) platforms, I would generally search for some recommended movie as I started my lunch and then begin watching as I had my lunch. More often than not, I would continue watching this movie after I’d finished my lunch and end up completing the movie in one sitting. On a few occasions I would also try out my luck with a series and in case I liked it, I would generally go on a weekend binge of 5–6 episodes (depending on the series). This was my simple content discovery journey & it seem to have been going fine for some time and I would hardly have a situation where I couldn’t find the right content during my weekend lunch. However, over the course of last 6–8 months something has changed. Although the volume of content across both Netflix and Amazon Prime have gone up, I now constantly find myself spending almost my entire lunch time searching for a movie or short series that I can watch — and on more occasions the search doesn’t yield a fruitful result. In a number of cases, I would just try some random movie and stop watching midway as I don’t find it to my taste. It seems somehow the famous Netflix recommendation has taken a hit (at least for my taste in entertainment).
What has changed?
Before getting into an analysis, let’s look at the evolution of the OTT platform in brief.
For a very long time since the advent of Television technology (and broadcasting), TV has served as a source of entertainment (storytelling through movies, series, etc.) and information (news, sports events, live events). The business model for broadcasting was based on Advertising. So consumers would watch their favorite TV programs for free while also watching native advertisements which ensured that the broadcasters and TV channels made money to continue providing more content. Advertisers were more than happy to spend huge budgets on TV advertisements to get their brands across to millions of viewers.
This business model continued for some time, until Cable TV came into the picture. With the arrival of cable tv in the mix, the business model went through some change. Now Networks and Channels had started to charge a fee from the cable TV providers (distributors) — who in turn passed this cost to the consumers. This cost was essentially an affiliate fee. This model had not changed even when the DTH (direct to home) operators started replacing the cable tv operators. Consumers kept subscribing to the services provided by the DTH operators just like they had subscribed to the cable tv operators. This change only ensured that operations could be consolidated from a lot of small cable tv operators to a few large DTH operators.
The breakthrough came with the realization that Internet’s almost zero cost of distribution could usher in a totally different model. Internet removed the means of distribution from the broadcasting value chain. So now the content creators/producers could directly provide the content to the consumers. YouTube was an early runner in this system. Soon OTT providers came into the picture to disrupt the traditional tv market. This saw the entry of Netflix and Amazon (and many others) into this market.
From a consumer’s perspective, the mode of access has remained the same i.e. monthly subscription fee. However, what changed is the mode of consumption
From broadcasting to on-demand viewing, from channel based choice to endless choice (at least as much content as an OTT platform offers)
However, with this choice comes the responsibility of finding out the content that you want to view.
Granted the recommendations provided by platforms like Netflix, etc. does help you to some extent, however as a viewer you would soon come to a situation where the search for the right content takes so much time that you would soon start to wonder whether you should watching something at all — this is especially true if you have slotted a specific time for tv viewing. This is because you start to wonder:
- if a particular series is worth 2–3 weeks (5–6 hours each week) of your time
- If a particular movie is worth spending 2 hours
- If you discover half way through a movie that you don’t like it
What is causing this problem of content discovery?
While there are many reasons contributing to this problem, the overarching problem is the content discovery experience. You see the OTT platforms especially Netflix and Amazon Prime have been pouring in billions of dollars in either original content creation or getting rights for existing popular content. Existing popular content is a way for such platforms to bring in customers who would prefer all their content viewing to happen in one place. Original content on the other hand is meant to ensure that such customers get hooked on to the platform.
Discovery of popular content is NOT a problem — because as a viewer you know what you want to watch and the platform would either have that content or not (search for such popular content provides a way for OTT platforms to understand which existing content they should get rights for). It’s a different ball game when it comes to original content. Original content takes the form of either feature films or series. The rate at which such original content is getting created is really high and you can get an idea about this if you start randomly browsing content on Netflix and follow the trail of “similar content” recommendations. It’s like going down Alice’s rabbit hole.
But the real problem is how do you choose?
To answer this, let’s look at how traditional content discovery happens:
Content Category: Movies — Discovery happens primarily through trailers people watch when they are in the theaters or on YouTube or as Advertisement. Viewers choose based on cast and whatever story can be discerned through the trailer. Another important discovery mechanism is word of mouth — however, this is really dependent on the quality of content. Movie reviews also play a big role in helping viewers choose content.
Content Category: Series — Discovery happens through trailers shown as advertisements during popular TV shows, YouTube, etc. Other methods like reviews and word of mouth spread also play an equally important role just like in the case of movies.
OTT platforms also follow the same approach when it comes to their hero content. However, that still doesn’t solve the problem for the huge number of original movies and series that are getting created. In the traditional broadcasting model, viewers would tune in to watch their favorite shows and TV networks could plug in new content between the time of 2 favorite shows and get at least some people hooked onto the new show (as long as the content was good). In the traditional broadcasting model, the volume of content needed was relatively low (relative to OTT) because the Networks controlled the streaming of content. The only choice viewers had was to switch channels. With the OTT format, viewers can watch the content they desire, however, they have to hunt for the right content (beyond the popular ones).
There are 5 things that play a pivotal role in this hunt process:
- Search
- Recommendations
- Trailers
- Reviews
- User Experience
Search
A large proportion of OTT content is viewed on mobile devices — which is where content search experience works well. However, search is really tedious on TV — especially when you use device like Amazon Firestick. Typing your query is difficult and voice search doesn’t work on Netflix. Even on smaller screens, based on the search results I have no way of understanding if the content results are based on relevance (for genre based search terms), recency or popularity. Also content that has already been viewed also appear on the top of the search results. Also there is no easy way to view any associated trailers without navigating specifically into a particular result. All of these make the search experience unimpressive. More importantly this also means that original contents remain at the bottom where discovery doesn’t happen.
Top 3 things that can be done to improve search:
- Search algorithm needs more personalization based on the individual search behavior
a) Don’t show already viewed content on top of search results
b) Don’t show content that user had started but didn’t complete for the last x weeks
c) Rather than sorting results based on just popularity (overall views), filter and sort based on views (and likes) of similar cohort
d) Bubble up original content on an experimental basis (at a customer level) and see if they bite
- Search result tuple changes
a) Allow users to view trailers (if they exist for any content) from the search tuple
b) Show ratings on the search tuple, it does help to get some affirmation
- Support voice search for non-mobile devices
Recommendations
The recommendations from Netflix works in a lot of instances, but there are instances when such recommendations fail terribly. I would be more than happy to let Netflix know about the movies that I have loved watching in the past through some sort of a form — but I need to be asked. Apart from using all the fancy recommendation algorithm, just find out my tastes explicitly. My viewing behavior on the platform might not really reveal what I want to watch, since it is limited by the content that is present — whereas I might have viewed content outside the platform which might have valuable insights into recommendations that will pleasantly surprise me.
Trailers
Trailers are so important in letting users decide if they want to view a content. However, a lot of content on Netflix and Amazon Prime doesn’t have trailers, so you are forced to take a chance and start watching a movie only to discovery 15–30 mins into the movie/first episode, that it really sucks. This is especially true for original content on Netflix. The OTT platforms can simply use machine learning to generate trailers from content. I know that Netflix surely has some form of this, however, I am not really sure why they won’t use this technology for most of their content — which would benefit the viewer. Whatever the reason, all OTT platforms should really invest in such a capability (at least for the original content they are producing).
Reviews
Getting reviews for watched content can be tough and generally a complete watch for a specific content is taken as an indicator for recommendation and search algorithm. However, here are a few things that should be considered before a complete view should be taken as a positive attribute:
- Is it possible that viewer completed viewing a movie, because they had already invested time in watching the first 1 hour (hoping it was a good movie)
- Did the viewer watch the movie in less than the overall duration of the movie i.e. did they fast forward through the movie
For complete views of a particular content OTT platforms can simply surface a thumbs up/thumbs down review towards the end of the movie (experiment with different formats) or provide some incentives to review movies people have watched. Needless to say reviews can be really helpful in search and recommendations and platforms should attempt to surface reviews in a seamless way without disturbing the viewer.
It is also important to bring in expert reviews to promote content. YouTube has content reviewers who have millions of followers. It’s important for OTT platforms to seamless integrate such expert reviews into their content or vice versa.
User Experience
Most audiences are lazy when it comes to finding the right content to view. It’s important that the user experience is able to navigate the user to right content in the shortest possible time, so that more time is spent viewing and less trying to search for content. Content discovery should be seamless and the user experience should blend in different aspects like recommendation, search, trailers, user interaction etc. to land the user on the right content which the user can then enjoy for the next 30 mins to 120 mins (not counting binge watches). This user journey should not take more than 5 mins.
Unfortunately for me the content discovery on OTT platforms (Netflix and Amazon prime) has started to resemble the experience of finding needle in a haystack taking up to 15–20 mins finding the right content. Beyond some of the points listed in this point, it is also because I have become much more choosy over the last few years when it comes to my entertainment content. Given I can spend only a stipulated time on content every weekend, I tend to focus on quality rather than taking the risk of playing the first recommendation that pops up. I am sure there are tons of user who fall in this category and Netflix and Amazons of the world should really figure out ways to surface out the right content in a seamless way. It’s not just about producing and creating lot of content/choices for viewers. After all too much choice is no choice at all.
