Effortless onboarding with a personalised reward
During open interviews where we aimed to understand the journey for first time users better, we observed that users peaked when they found a video they wanted to try, but had to put in too much effort to get to that point. We wanted to create a new onboarding that helps the user to get to a guaranteed peak effortlessly.
The Team
- 2 Designers
- 1 Product Owner
- 6 Developer
- 1 Marketing consultant
- 1 Content consultant
My Role
- Market Research
- User Research
- Usability Testing
- Wireframing
- Prototyping
- Workshop Facilitation
Used Tools
- Figma
- Miro
- Rapid Usertests
- Jira & Confluence
Duration
4 Month
Problem
The pandemic made online yoga boom and a dip in new users signing up and subscribing was expected, but we observed that the dip in bought subscriptions was more drastic than predicted, especially on mobile.
Goal
Get more mobile users to purchase a subscription after signing up and completing the trial period.
Challenge
Identify the most urgent reason that makes users decide against buying a subscription so we can come up with a solution that will lift the dip.
The Process
Open interviews
... to observe new users experiencing our app for the first time so we can learn where problems and doubt accrue but also where users find value.
Ideation Workshop
... to gather ideas that might solve the most important problems we found and provide users more quickly with the value we identified. The idea we chose to place our bet on was a new onboarding that would end in a personalized video recommendation.
Market Research
... to understand existing patterns of onboarding experiences.
Testing Competitors
... to learn which approaches work well and which we should avoid as well as learning about what users expect from an onboarding and recommendations in general. It also served as a validation to our hypothesis that this solution would help solve our problem.
Internal Interviews
... to inform our onboarding concept. Yoga teachers helped to define how to ask which questions to give valuable recommendations while the content team helped to match answer options with meta data.
Test with Wizard of Oz Prototypes
... to learn how well the users would get through the onboarding and what makes a good recommendation.
Iterate and Test
... with users again and again till we identified and solved all usability issues and refined the concept that we want to develop.
Technical Proof of Concept
... to identify and understand possible feasibility issues and tackle big tasks as early as possible.
Adopt Concept to all Platforms
... to also help and serve recommendations to users that sign up via web or tablet.
User Journey from Open Interviews
Faking a recommendation engine in Figma
We wanted to achieve...
Flexible journey
Users should not be forced to do the onboarding if all they want is to search for videos on their own. Instead they should be guided to exactly that option quickly.
Flexible reward
We found that users did not enjoy a list of videos that was set in stone, but where also not always willing to create one on their own from a preselection. We ended on offering the choice to take it as is or edit it right away or even later on.
Temporary
Users wanted something for now but knew that tomorrow their needs might change. We made sure that it is clear from the beginning, that answers will not feed a permanent algorithm.
The right amount
After experimenting, 5 videos seemed like a good amount of videos to not over or underwhelm.
So we decided on...
It has to be personalized
We saw that users latch on as soon as they find a video that suits their current needs. As such we included questions that ask for those and used the gathered metadata to fetch video matches.
BUT we found that not all metadata is equally important. As long as time, style and name of the video seemed (!) to match their preferences, the video would speak to the user enough to make the decision to watch it.
The reward should last
To make sure users would come back the next day, we wanted to give them not only a feed to browse, but a personalized package (list of selected videos) they are able to explore within their trial period. By making them edit the list and give it a name, an emotional bond was created which made it more likley that the user would actually make use of their trial period.
Onboarding Questions
Recommendation Selection
Figma Prototype - Onboarding Questions 1
Figma Prototype - Onboarding Questions 2
Figma Prototype - Onboarding Questions 3
Figma Prototype - Recommendation Selection
Fun Fact
In the click dummies we faked the algorithm so convincingly, that users where actually interested in the videos they got and signed up after the test.