COVID has made the world look like a riskier place to some of our offer partners. Our users were seeing less products and our goal was to help reduce lender risk by providing new live data and insights so they could make more informed credit decisions.
The pace of delivery was going to be quicker than any project we’d set out on before and we wanted to launch on all 3 platforms (Web, iOS & Android) simultaneously.
This work would also be highly dependant on partner execution and commercial agreements.
Our main objective was to connect 100k user bank accounts by the end of the race. The results we’d target to achieve that were:
OB seen: 40%
Of users who logged in, what % saw an OB prompt?
OB clicked: 30%
Of users who saw an OB prompt, what % clicked to connect?
OB linked: 24%
Of users who clicked to connect their account, what % completed the link flow?
OB pulled: 65%
Of users who completed linking, what % successfully pulled transation/report data?
We spoke with a few people within the business who had experience working with OB and downloaded several OB powered apps to go through the connection journeys. We had to move as quickly as possible, so with this limited research and understanding of the problem area, we went ahead and put together a quick prototype.
Typically we have users come to our office every week and we can easily book time with them for one-on-one sessions. As a result of COVID, we would have to conduct our testing remotely, using a script and the prototype I created in Marvel (both uploaded to UserTesting). After screening we ended up testing with 5 users. We watched all the videos back while taking notes in Miro.
The insights were really useful. We noted that people had a lot of anxiety around linking their accounts and weren't too happy with this experience. We had our work cut out for us!
We tested an entry point on the offers list page but had to rethink this when a potential partner required the OB data before a users sees a non OB offer.
We designed an entry point on our product picker and loan questions page:
If we don’t identify which bank you use from the credit report, you can select it from a list of our supported banks. When we designed this screen we were surprised to find many more supported banks than we’d initially thought there would be – a bank might have multiple accounts as well (ie personal and business). Our MVP ended up having a long list and not a great experience finding user banks. For phase 2 I designed it with a ‘type to filter’ input and grouped the bank account types to the bank. In hindsight, I should have got the exact list from the API a lot earlier.
We ended up going live on all platforms on time, a massive shout out to the dev team for that. However, we didn’t hit our 100k user target, as we ended up changing our OB connections partner. This change meant that our dev team would have to update our backend before we could think about scaling up.
We built the MVP to a great standard considering the timeframe and scope of work. I’m looking forward to seeing how else we can improve offers for our users with OB data such as including income verification and affordability insights.