The dashboard personalisation project that improved engagement by 18% for PreK users.
Problem statement
At SplashLearn, my first project after transitioning into product design was to improve conversion rates for lower grades (PreK and Kindergarten) which were extremely low compared to the higher grades. Additionally, the share of younger kids within the total user base was significant, which meant this segment represented a large opportunity for growth.
SplashLearn is a learning app for the US audience. It offers gamified learning content for PreK to Grade 5 (Math and English). The app offers a 7-day free trial, during which parents and kids can have unlimited access to all the content before they are credited for their chosen subscription plan.
Process
Relooking at the user flow
We have a relatively long user flow that includes parent and child onboarding culminating at the free trial offer. The diverging flows here result in the two users:
Freemium: The users who opt of of this offer are directed to the freemium or the limited offering. They can upgrade at any time using the 'Unlock all' CTA and start a trial from the child dashboard.
Premium: The users who start trial are presented with the premium product and can cancel within the 7 days. Payment is deducted on Day 8, after which the user is considered a successfully converted user.
In both cases, users see the dashboard every time they open the app (logged in), and thus, the onus of paid conversion heavily relied on this initial experience of the product. Our key metric for this project was T2P.
Interviewing our users
We presented Pre-K parents (potential customers) with the post-install flow to understand their concerns and derive insights. The following issues were identified:
Content Discoverability: Parents struggled to find relevant content tailored to their preschoolers. While our search data showed certain popular keywords (Colouring, Letter Tracing, Shapes, etc.) some of these were either not readily available or not easily discoverable within the app, making it difficult for parents to find what they were looking for on Day 1.
Low Parent Confidence: Parents felt uncertain about whether the app provided the content they needed for their children. This lack of clarity was preventing parents from moving forward with the trial-to-paid conversion process.
Finding the gaps in our marketing promises
Appstore:
The Appstore is optimized to feature our app for certain keywords and also includes Custom Product Pages (CPP). The keywords and features highlighted in them are derived from data insights. The obvious next step was to tally if the user journey encountered what we promised here.
Child dashboard:
Our primary view of the existing dashboard focused on a daily learning path and exhibiting curriculum coverage. While this was age-based, the view was more suited for Grades 1-5 than preschoolers. The most searched keywords that SplashLearn offered for this age didn't nest in this view.
Approach and implementation
Hypothesis: Easy-to-discover and need-based content on Day 0 would encourage parents to engage better and take a positive conversion action like trial start and reduce cancellations.
We believed that if the parent is easily able to find need-based content within the first view of the dashboard, they will have much more confidence about SL providing the content that they are looking for.
The solution centered around personalizing the user experience using two key insights/data points:
Dashboard rails using broad trend data:
We used this data to adjust our dashboard's information architecture, expanding our categories beyond Math & English into skills, such as Motor skills, Creativity, Problem-solving etc. This ensured that popular keywords like "Colouring," "Tracing," and "Puzzles" appeared more prominently in the dashboard for all users in the relevant age group.
The thumbnails for each card was also re-created to convey the learning skills (earlier game screenshots were being used)
Dynamic content using child assessment:
We recently introduced a brief child assessment in the onboarding flow to understand each child's proficiency in key learning areas (Math and English).
Based on these results, and whether or not the subjects were personalized, the dashboard would order and display content. For personalized users, activities and content that were most relevant to the child's developmental level; for example, if a child struggles with letter recognition, the app will surface letter tracing or sight word activities on the Language rail.
While the above solutions tackled the problem metric (Trial to paid) directly, we also looked if we could enable more parents to start the trial. Without elongating the onboarding anymore, we found opportunities (loader placements) to reinforce that their search-relevant learning content lies beyond the paywall.
Team
The project was led by one project manager, one product designer, one game-art designer and two developers (web and iOs).
Outcome
Both projects were tested in combination and independently (only dashboard) against the control variant and showed improvement in our conversion metrics. Currently they are at a 100% rollout and used in combination.
Key Takeaways
Personalization Drives Engagement: Tailoring the app experience to the specific needs and behaviors of users is critical, especially for early education where parents are involved in the decision-making process.
Content Visibility Matters: Ensuring that the most searched and requested content is easy to find on the dashboard can have a direct impact on user retention and conversion.
This project was highly data-driven, deriving insights from the highest searches and content engagement - helping in cutting out the noise and highlighting what matters most for an intuitive and personalized experience.
Comments