User Retention Part-I

Defining User Retention

User retention is difficult to understand and easy to bluff. Due to the lack of common definition, it can be easily misunderstood. It varies a lot with product and context so no single size fit for all. In this 3 part article, I will try to explain what and how user retention should be measured for Video streaming platform.

Part-I – Defining the User Retention

Part-II – Metrics to measure User Retention

Part-III – Common Mistakes to avoid in measuring.

“It is my opinion and happy to get corrected with some wise comments”

Loosely,

Retention rate is the percentage of users who use your product recurrently over a defined period of time.

Why did I say loosely?

Because it can not be defined without having context.

So, In the context of OTT (Video streaming),

Retention rate is the percentage of users who use your video streaming service multiple time in the given time frame/window with respect to the defined event frame.

For some, it might be the simple definition for other it may be complex so let’s break it down.

Use your video streaming

Defining the app usage is varied across the platform but every product has a core objective to achieve. For eg. e-commerce want to sell something, so they might track user purchase while some other product intended to maximize their subscription. But, most of the companies choose the biggest number or number which they can influence or control (App open, Video Impression, etc.) to define their retention rate. On a similar line, many OTT player uses two metrics to calculate their retention rate.

1. Impression: It’s an event which occurs the moment you see thumbnail or the first frame of the video in the player.
2. Interactivity: It’s an event which occurs when you interact with the player, like press Play, pause, seek, etc.

Why?

1. The bigger number better is the investment
2. You can influence this number tactically by external stimuli like Push Notification, Email campaign, etc.

So, with a few dollar investments, you can drive the user to open the app and get your desire retention rate. The problem with this approach is by neglecting core product objective in driving growth, you can’t have a solid product strategy. And without the great product strategy, you can not maintain that retention rate over a longer period of time. There are other demerits as well, but let’s keep that for some other time.

So the question is, what OTT player should do is, include the third factor while calculating app usage.

Engagement: Although it might start with interactivity, it is way too complex than that. In the current context simply consider it as the duration of video played.

So, what is the challenge in considering engagement?

1. It will be the smallest number. No one like it 
2. There is no single way you can define engagement. What if one considers 5 second and others may consider 10 seconds? What if someone wants to define it like a 90% completion rate? It’s relative so difficult to define and agreed upon easily.
3. Watching video is one type of engagement but browsing and exploring through the app could be another kind of engagement. (Maybe he want to download). This might not be so relevant to SVOD model but what about AVOD?

So simply relying on video watch is also not feasible. The proposition here would be to consider both and assigned weight as per your product strategy. For eg.

Let’s suppose give it a higher weighting (100%) while calculating retention rate whereas give low weight (60%) to the users who are exploring the app.

How it helps, well for movies it will be more on watch side and less on exploring side, but on short video or Video songs, it will be less on watch time but more on exploration. So balancing out the usage percentage to conder is equally important.

 multiple time in the given time frame/window

It’s relatively easier than defining the engagement but still one needs to be very careful about how to define it. A question like, Do you want to consider a single window or you multiple time windows is required to give an accurate picture? Companies like Hotstar might have a higher retention rate at the time of IPL but what about the rest of the year? What if they consider the yearly retention rate? User will come to there platform but does it reflect correct retention rate? How about their daily and Weekly retention metrics? To overcome this kind of confusion one should use multiple time window like, daily, weekly, monthly, yearly, quarterly, etc.

But this might also not reflect correct retention rate sometime. For eg, While GoT launch in India, daily and weekly may go higher but quarterly might not change. So what window one should choose. There is no right fit for all and depend on your core business. If you are in web series, monthly will be a good option, but if you are in short video or Video songs, daily will give much clear picture. If its mix of all, assign some weight and come up with some useful formulas rather then simply boasting numbers. But having said that, defining retention in time required frame of reference, which can be provided by our third part of the definition.

with respect to the defined event frame.

Even time can not be measured in a single frame of reference so defining the reference is important. To understand this let suppose one of your favorite OTT players have a high engagement rate as it produces awesome web series. For some reason, they started loosing on this front and to entertain the user, they launch new categories (Let say Video Songs). With the right product design and some marketing gimmicks, they might engage the user in this new category but now, the user behavior is completely changed. They are now more frequent to the app, given short form of content and there could be a sudden rise in the daily app open. What if they change their retention based on app open then where they started? Without knowing the reference which events are being considered for retention can be dangerous.

So retention rate needs to be carefully evaluated for the parameter that is factored in for different time frame with reference to events. Best is to have multiple retention rates and give weight to each to reach a common retention rate.

So next time, when some OTT player claim there retention rate is 97%, just ask them in what time frame and with reference to what event. May be you might earn free subscription