Read and write movie reviews in 10 seconds (or less)
How BookMyShow improved the user experience for millions of movie watchers per month
Overview
Ratings and Reviews are an integral part of making a buying decision about any product. In the entertainment industry, they are more so relevant because of the short life of content like cinema movies.
A movie releasing in India on Friday has an average timeline of a few weeks to run in cinemas and generate profits. In that timeframe, if a movie gets good reviews, then it is likely that more people and go and watch it, vice-versa.
How might we revamp the rating experience so that —
• users take informed decisions about watching a movie
• users get an optimal experience while entering a review
This case study documents our experience building a very well received product — both by our users and the company. The outcome was a reduction of the time on task from 60 seconds to 10 seconds, and 24% growth in reviews.
Team: Jasdev Singh & Yashodhar Upadhyay
Role: Data analysis, Strategy, Research, Prototyping, Visual Design, MVP planning, User stories
With more ratings comes more responsibility
- Every month over 18 to 20 million tickets are booked by several millions of users on average. Out of those, let’s assume about 100 people rated a movie, but only 8 of them wrote reviews.
- BookMyShow’s strong brand value ensures that a movie that gets good ratings, is likely to see more users buy tickets for the same, and vice-versa.
Houston, we have some problems
1. Design inconsistency
At one time there were over 10 variations of the heart & star icon being used on the platform to denote ratings.
2. User contribution not highlighted enough
Someone might have taken inspiration from Twitter and brought a minimum 140 character review limit, but our users were using unrelated text, repeated words, or emojis to get over this mandate. And, for those who took time out to write proper reviews, their contribution was going right down the abyss.
3. Few direct nudges
Users could only rate a film from the movie details page, or from the homepage popup modal. For instance, if a user opens the app after watching a movie, there were no additional touch-points to help users rate.
4. Lack of trustworthiness
Anyone could come and add a rating or review. A person reading reviews would not be able to ascertain if the writer had actually seen the movie or not.
Research is like a box of chocolates — you never know what you’re gonna find
Crunching data
We looked at Google Analytics and Clevertap reports.
Our existing rating form UI had a big problem — the rating scale was from 0% to 100% in 10 point increments, but the UI showed only 5 hearts to capture a rating. That meant there was a case when only 1/2 a heart would be selected. The UI didn’t give that affordance, and data confirmed it. Because of the faulty UI, users were entering, for example, a 4 rating more number of times than a 3.5 rating.
Secondly, we found that for every 100 people who came to rate, only 8 of them wrote reviews. This was interesting as one of our goals was to change this behaviour.
To deep dive into this we next moved on to find more about the mental model of our users via qualitative research.
Surveys, phone calls with people
The first one was done with 62 users via a survey (and phone calls with some). The second one was done with over 1400 users through a survey. The audience was selected as a subset from our existing BookMyShow personas.
“I’ve never written a movie review in my life — I’m not sure what to write, and even if I do, I don’t really know if anyone’s gonna get value from it.”
— A user
Questions for which we wanted answers:
What influences users to write? — 60% write when the movie is either too good or too bad, 18% write because of intrinsic motivation, 14% write to help others
How do they make buying decisions for a movie? — Trust their friend’s word of mouth, search for ratings on IMDB, social media and BookMyShow
Why do they not write reviews? — Writing is time-consuming, Don’t know what to write — writer's block
Common phrases in text reviews
After looking at more than 300 distributed reviews, an interesting discovery was on the content — that users were writing. In the Indian context, there are a few repeated terms or jargons that cine-goers use to express emotions for a movie, for eg: Jhakaas, Blockbuster, Must watch, and similar such.
Looking at what our friends are doing
We studied the end to end rating and review flows of various other apps to look for commonalities, useful patterns and benchmarking optimal flows.
Does a one ‘size’ of rating fit all types of use cases?
Summarised findings:
- Apps like Zomato, IMDB have a graph like UI where they show the breakdown of rating values
- Uber has a concept of driver compliments
- All platforms take ratings on a 5 point scale, with the exception of IMDB, which has a 10 point scale
- Instagram has a slider which users can place in stories and get other users to engage with it in a seamless manner
Ideate fast, fail fast, learn fast
We then brainstormed both individually and as a team, for ideas that we could implement to solve for the different user pain points. We didn’t limit ourselves with any existing tech or product baggage. The 30+ ideas were then weighted on an impact vs effort graph and few key ideas were prioritised to be designed for release.
Heart vs Star
Our branding team had finalised the heart construct for ratings, which is not an industry recognised standard for ratings which typically uses stars for ratings. We had raised this point with them as a usability concern, and the trade-off here was to go with the heart construct.
Adding more love to the heart
Existing mechanism: 5 hearts with 1/2 heart increments. No affordance and our quantitative data confirmed it.
Take 1: 10 hearts are more complicated than just 5 hearts (Hick’s Law: reduce the number of options to reduce cognitive load and improve the speed of decision making).
Take 2: Range slider. Something playful and easy to use. During tests, we found that it doesn’t have drag handle affordance.
Final take: 10 point slider with snap affordance, drag handle affordance, and haptic feedback. Performed the best in our tests.
Doing the PM stuff: New touch-points and user stories
We stressed on new touchpoints (eg: push notifications) across the platform where the user could be nudged to submit their rating. This is where we again stepped into the shoes of a project manager and collaborated with our Agile coach, Chhavi Raina, to learn how user stories are made in a GIVEN, WHEN, THEN format.
Build once, use anywhere — Rating and review input form
We tested the below iterations with users and found the last one to perform the best. It used the progressive disclosure model of showing information in a 3 step process — ask for a rating on the slider, ask for hashtags (optional) and ask for a written review (optional). The hashtags were derived from the content of reviews, as we saw in our research.
The time on task for adding a hashtag review was as low as 9 seconds! It provided the right balance between asking all 3 questions and keeping the UI minimal. Plus, it was a modular component which could be used anywhere on the platform.
Tradeoff: Based on the research we found 88% users included cinema experience in their ratings —and in some variations (such as above), we can see the cinema experience question. We decided against including it in the initial launch since it required the whole business loop to be thought through— How would we use these ratings, and how will the handshake with cinemas look like?
The “details” on the movie details page
This page displays details about a movie. We followed the mental model of a typical product detail page where reviews are usually placed at the lower scroll of the page. The top 5 hashtags and the top 5 reviews for the movie are shown in the reviews section, the combination of which is a very powerful way to show a summary of user sentiments. During tests, users were able to skim through in 10 to 20 seconds.
As for additional touchpoints to nudge the user, we added push notifications and nudges on the homepage.
Increasing trust and combating fake reviews
To combat users’ low trust in reviews, adding a simple “Booked on BookMyShow” text alongside each ticket-holders’ review helped significantly. The rating algorithm was also tweaked to give more weightage to these kinds of ratings and reviews while counting the aggregate rating of a movie.
To solve fraud, we strengthened our backend fraud detection algorithms and the system can now observe fishy patterns and flag them to admins.
Final_final_final_design
The happy flow looks like this. Enter your rating and review in about 9 seconds, plus a delightful animation at the end! ️
Try it out on the new BookMyShow app 😍
Key results
Amazon India partnered with us to list movies on their app and Ratings are an integral part of that handshake. This also boosted the number of mentions and shares across social media.
Insights
- People don’t type — it helps to prompt them with nudges to get started
- People don’t read — they want TL;DR versions of the text
- PMs have a tough job coordinating with everyone. Respect ++ for them!
Key members: Kedar Nimkar, Sandeep Nair (Design), Akshansh Dhing, Mohit Kejriwal, Tanvi Nabar, Pradip Shinghare (Front-End), Apoorv Kumar, Chakshu Ahuja, Sneha Dhumal, Rinkal Kevat (Back-End), Punarvasu Pandey (QA), Aditya Kuber, Manvi Sharma.
