At Skrill, our registration and onboarding process was sub-optimal. We identified this by looking at conversion rate numbers and doing a series of user interviews.
Our goal was to decrease the drop-off rate in our onboarding process.
Our onboarding was burdened by technical and UX debt. We knew things could be improved and were further convinced by looking at the data – there was a big drop-off when people were asked to provide additional details and complete their account.
To gather a complete picture, we conducted a series of user interviews. What people told us matched what the funnel data was showing, providing more detail and nuance to the problem.
We discovered that the user journey was very fragmented, did not follow a predictable and logical flow – this frustrated our users.
Our digital wallet provided many value props, each of which had its little stack prerequisites and flows. Instead of opting for a complete overhaul, we decided to go for the least amount of changes that could generate the most improvement and could be tested and observed – an MVP, if you will.
Looking at the data, we noted 2 important issues:
Issue 1: We did not do a good job of building a mental model for our users – our light dashboard gave the impression you could do many things, while in reality the only thing you could do is complete your account and make a deposit.
Issue 2: Increased context switching and deviations from their main task confused our users.
The idea to improve this was to reduce context switching – tweak people's perception that every step, as well as their first deposit, is part of one single 'account setup' phase.
We could achieve this by setting expectations at the beginning – what is possible and what is necessary to do the thing I'm here to do...
... and make our system provide feedback at each step of the journey - where the user is and what's next.
Beyond these optimisations, the idea evolved into capturing the user's intent at the beginning, providing relevant information and guiding people beyond their account set-up to their main goal – what the person wants to do after they're all set.
Capturing intent at the beginning enables us to point the user to their desired path at the end of the onboarding journey:
A few tests were done on usertesting.com with Figma prototypes prior to development. The goal was to gather perception and validate there are no major gaps in our concept.
However, due to the fiscal nature of our product, we wanted to perform live tests with our customers at scale. We did this using Adobe Target in a few key regions – approximately half of the people were shown the new version of the onboarding and the other half the current implementation.
In the spirit of fast and lean iteration, we leveraged our existing design systems and UI patterns to minimize development time and try to validate concepts instead of visuals.
We saw an improvement across all areas and regions that we tested.
In some parts of the funnel, the drop off rate was decreased by close to 30%.
People were completing their account registrations more often.
In others areas the change was not that big – deposits in the UK, for example, saw only a 2% improvement.
This experience forced us to look outside of our silos and consider flows more holistically – people don't want to just set up an account and deposit money for the sake of it – they want to do something else outside of our product and our features are just means to an end.