top of page
A foray into personalization
Designing the new For you page
A one-stop hyper-personalized “For you” page that helps users discover relevant content quickly while keeping them engaged.
Organisation
ZEE5
Role
Product designer
Timeline
10 Weeks
Responsibilities
Design direction
Research
Data analysis
Interaction design
Visual design
Prototyping
Developer handoff
Overview
The story did not however begin with a brief to redesign the new For you page. I was approached by our PM, with a different ask.
" Every year ZEE5 releases some high profile or marquee content, these releases drive subscriptions that would be lapsing in the coming year post purchase of the subscription. Why don't we show a pop up to these users, 30 days before their pack expires and nudge them to renew their plans with personalized content recommendations "
While the idea sounded good in theory, there were also some concerns we had and we wanted to challenge the brief to understand what kind of impact this would actually lead to.
Initial challenge
The question on my mind was, will it really be effective in converting some users, if we nudge subscribed users 30 days before their plan ends to get them to renew their subscription?
To validate this, I had multiple rounds of discussions with the UX research and Data Science teams to collect insights that could help us understand this problem better.
Systhesis of discussions with research
The insights received from research were that There is a pattern which exists for users to buy plans for Marquee content pieces. There was however another behaviour noted that if there was only one appealing piece of content then the tendency to borrow a plan became higher.
Meeting with stakeholders
We went back to our stakeholders with the insights we received. We stated the following problems to them
A nudge 30 days before the end of a subscribed users' plan will target a very small number of our users.
If there is a problem with recommendations earlier on due to which there is a significant churn, then it would be more difficult to get this right at such a later stage in the user's journey with our platform.
Synthesis of discussions with data science
There is an increase in subscriptions during the release of marquee content. There is however a gradual churn from the platform for these subscribers. By the last month there is a sharp drop in the number of returning subscribers to the platform.
Our proposal instead
" Why not create a personal space for the user that is just theirs and provides tailored recommendations from the get go "
our proposal was accepted and the "For you" page was selected to be revamped for this purpose. The task at hand was to create several new features that could make this page highly personalized and overtime replace the ZEE5 homepage after a few versions.
Redesigning For you
High Level Goals
Increasing the viewer to user ratio (V/U)
01
Increasing the number of video views per viewer (VV/V)
02
Initiation
I led the product design and mentored an associate designer on this project who collaborated with me on this project. We worked with a cross-functional team including a Product manager, 11 Engineers (front end, backend and QA), 1 UX researcher 1 UX copywriter and 2 data scientists.
Diving into research
Research methods used were secondary research from sources across the internet on personalisation as well as analysis of the foundational research conducted on our users by the research team. The researches were divided into two cateogires, AVOD (Advertising video on demand) which meant Freemium users, and SVOD (Subscriber video on demand) which meant subscribed users that paid for our subscription to watch content ad-free.
HOME PAGE
CURRENT
FOR YOU PAGE
Pain points with the current experience
Starting with an analysis of the current design of For you. there is no difference in the current design and features on For you.
LIMITED CONTENT RELEVANCE
The current "For You" section seems more like "For Everyone"
The content available is not personalised
NO DIFFERENTIATING EXPERIENCES
No hook for users to return
CONTENT DISCOVERABILITY ISSUES
Discovery is not simple
Uncurated/repeated content causes decision fatigue
The journey to finalize our feature list
To reach a viable list of features, we used the design process to first narrow down to the list of features we wanted before we started wireframing.
The core steps we followed are depicted here but this process was not a linear process, there was a lot of back and forth between steps and discussions with stakeholders before we reach an acceptable level for each step.
Collecting attributes that help make watch decisions
This was done to help understand the hierarchy of information that would help our users use to make a watching decision.
Benchmarking of OTT and non OTT based apps for personalization
We did a thorough benchmarking of different applications to help us understand successes of other players
Collecting all available rail titles and categorising them
This was done to avoid doubling of personalized rails and to also understand modifiable components to reduce dev effort
Using personas created by the research team to narrow down our TGs
We used personas created by the research team to help us narrow down who we were designing for.
Working with the product team to create user scenarios.
We used these user scenarios to draft HMW statements that helped us understand personalisation problems we could solve.
Brainstorming a feature list
We then brainstormed a feature list for each of those personas, looked at common emerging themes and took that list to product and Dev to understand effort.
We went through a process of creating several iterations of our ideas and combined with several validation calls with senior design team members and product team, we landed on the final screens and features.
Wireframes
We went through a process of creating several iterations of our ideas and combined with several validation calls with senior design team members and product team, we landed on the final screens and features.
Wireframes
We created a demo video for the phase-I build of the page
Welcome to the new "For you"
experience
Carousel Banner
MAIN BANNER
Large cards added for focus on content recommendation with condensed details and tags to nudge decision making.
CUMULATIVE RATING
Cumulative rating populated by AI model from rating websites. as a generative AI used case for social proofing
TOGGLE SWITCH
To nudge AVOD users to get a glimpse of what an SVOD user sees.
Because you watched
Match percentage is based on relevance to both the content tags as well as user watch behaviour on the platform
SOCIAL SENTIMENT
Cumulative rating populated by AI model from rating websites.
MATCH PERCENTAGE
First OTT platform in the industry to have generative AI used cases
Using UX persuasion methods where social proofing a commonly used method for persuading users to perform tasks. Our team used Bard's capabilities to sift through the internet and provide one line positive social sentiments that will be displayed in different rails in different iterations.
Social sentiments
Multiple stoylines
Our TV show audience is usually playing catch up a lot with the latest episodes of their favourite TV shows so we thought why not keep them more engaged and Gamifying user engagement for TV shows and web series
CREATING A GAMIFIED EXPERIENCE WITH AI
Creating engagement for TV show watching audience through gamification.
HYPOTHESIS
People like to guess and predict and have the desire to know if their prediction was correct.
Introducing entry points for cross discovery
ZEE5 SHORT CLIPZ
ClipZ was the name bestowed on our new reels format by our UX writing team. We designed an entry point for this format from the "For you" page.
AI SEARCH
We also created a custom entry point into search. This was added as the number of impressions fall as users start scrolling vertically, so this section would pop up before a certain number of rails base don the user's behavior.
The new For you page was launched to 10% of user base for Android and iOS as an A/B test for two weeks.
A/B test
V/U ratio up by
57%
Post two weeks a consistent spike in 8% was observed. Every 1% spike in v/U ration increase leads to an incremental revenue increase 2Cr.
Testimonials
The project garnered recognition across teams with positive feedback from senior management and won the internal hackathon as well.
bottom of page