To aid in the experience of personalised learning outcomes we noted down e specific and personalised learning improvement programs for each user cohort. We duly noticed the change in content consumption cohort based on cohorts and sketched out a college prediction engine that we will discuss ahead.
Our final results help account academic weaknesses, behavioural irregularities and gaps in test taking strategy that are holding students back from a seat in their dream college. Sounds interesting yet?
With this Embibe takes college prediction to the next level, as good as JEE counselling, with personalised academic revision plans and specialised learning frameworks.
Rankup has improved scores of lakhs of students by upto 60% in average.
Research and analysis
In a competitive exam like JEE or BITSAT marks to rank have a very unfair equation, for tens of marks lost a student ends up losing 1000 ranks. So we got together a team of students, data scientists, educators, engineers and designers to solve this problem through analysis and data. We decide to aim at a staggering 60% improvement with 2 to 3 tests taken for an average student, improving their confidence and morale along with it.
Interviews with stakeholders and students drive us to the understanding that every question is independent and it is about learning to always deliver your best, with each question. To enable and achieve this the cohorts we have made from our previous Research and redesign project help us group users into scenarios depending on their end goal and the current question. We note these cohorts down as Fighter, Performer, Achiever and Ranker from not so great students to the best ones in that order. Each user can progress to the next cohort easily by using Embibe's data science to their advantage and each user has its own preferred pack to unlock. The lifecycle of the product is growth based for every user and the product becomes a richer experience for the user as it is used more.
We figure most students worry about what to attempt first, how to fix a careless mistake, where time is wasted, what to prepare the day before to achieve an increase in 20% to 30% of his marks. In this chaos we need to make sure every student secures every mark he/she is capable of, that's quite a daunting task even when the students know all the answers. We come up with question packs for consumption, based on attempt data and consumption trends of the cohort last few years.
Scoring is therefore a behavioural game, it depends on your cohort, on your habits, on what you do every time when you see a question and Rank up is going to make visible this behavioural strategy to the students in a learning curve. We also position an effort based analysis and strategy thinking to improve test taking. This along with the Selected topic improvements feature is to satisfy the need for focused attention to each students approach depending on the exam scenario. Add to this effort rating (time and effort spent on questions) for every test to predict a score improvement pattern that we attribute to the wholesomeness of connect between user data, system cycles and the attempts based data.
These are called plans and the provision is made such that a unique plan is made every time the user can recalibrate Rank up and it calibrates to your dream college showing you exactly where your last 3 Rank up attempts can take you.
Our core principles are to narrow down the scope, make you work hard and predict where you are going to be all the way through your 60% score improvement plan, Ranking you up as you go for every attempt, with every query answered, suggestion and tactic shared. Throughout the lifetime of the build we locate key stakeholders in management, teaching and coaching institutes, educators and schools to fix and grow the product for maximum output for their own students.
Our Head of Data Science comes up with the cohort model and each team accordingly sets principles and practice in place to collect, analyse and treat data as was the goal.
Creation and framing
This project is no small task, we need to,
We start by building this system into the existing design and plan to come up with the new product as we make iterations on this design. We are aware that the property needs to be specifically different from the entire of Embibe, as it is a premium product tailored to make each individual student better. At the beginning we introduce packs into Embibe and a way to generate and buy them from the Embium store.
At this point students can consume these packs on the Learn and Practice properties and these packs refresh themselves with every attempt taken. They come at 400 to 1200 Embiums for each pack, depending on the level questions, guidance and personalisation it comes with. To clarify Embiums are Embibe's in-app currency which can be bought with real money. The initial UX design was just incremental to see the kind of consumption patterns for these packs and how students interact with these packs. Along with this we have a temporary Score++ landing page where students can sign up to the program specifically, for when it comes out.
Our suspicions prove right as 2 month into this update we see heavy traffic and consumption of this property, especially even more so during exam time. Meanwhile we begin to note down the second phase of this project, now that we have gauged the demand and the consumption patterns, we were to spend the next 2 months designing, developing and testing out the UX of such a product.
We start building this product out separately under the property "Score++", we add special packs according to the data we have, fix issues in the previous flow like payment and storage, spread interlinking across other properties and finally expand out the Score++ product. At this part of testing we decide to involve student and teacher stakeholders considering the fragile nature of this project.
The first round of screens
Score++ is a set of 5 recursive steps based on our insights and knowledge as an education team to improve your learning efficiency in detail,
So instead of having a one time personalised revision students can do it again and again to push their scores higher, while each attempt brings back fresh content, data and suggestions. They take a test,
With the launch of Score++ our first monetisation test efforts are underway. The first couple of months see good amount of users unlocking and consuming these packs. We have a financial target to reach with the trial run so we can be wary of the scale of consumption of such a feature.
To certify this premium feature and understand the use cases of such a product we go around the country to the best colleges, teachers and coaching institutes to test out Score++. We collect useful material from the stakeholders that you will see implemented in the following design of this premium product, RankUp.
The final design output
We chalked out that the key features of RankUp need the following systems design to be in place,
Needless to say that this premium feature needs it's own property not just as a part of the other Embibe properties such as Learn, Practice and Test, but a guaranteed rank improvement program. The core product mimics the the UX of Score++ as you will see ahead.
As you can see on the UX for RankUp the user is to first land and assess their test-taking skills, after which they are given improvement strategies and values. Once the user has taken more than 1 test the summary updates itself with the test taking graph for the user through the last attempts, along with cues to what to attempt next.
The purpose of the previously mentioned case study is to put the design project into context and to show how it works in a real world situation, as opposed to showing only how the final product looks.