Nubank's personal loan is one of the projects I'm most proud of. In this 10-month project I was responsible for product design and UX research (we didn't have a UX researcher at the time).
It was a quote-unquote “textbook design project”. An example of the double-diamond in use.
That's why I structured this case-study in “discover & define” and “design & deliver”.
Day one of the project and noticed we didn't know much about personal loans. So we went after market reports. Thing is we found only one. It wasn’t much, but enough to have an idea of the market landscape.
gets the one with the fastest approval
don’t have time to do research
look for instalments that fits their income
We needed to check that info and what the report didn’t inform us or what was different between our customers and the general public. To make things faster and cheaper, we focused the research in our customer base.
The majority hire at their own bank
The majority research in 1.5 institutions
Now with the surveys responses and analysis, we wanted to understand the why's, so we invited 14 people that answered the survey to come to the office to talk to us.
“I can’t understand why they treated me so bad? I went there to pay it.”
“I asked for 1000 reais and they charged me 1200.It’s an absurd! If I asked for 1000, I should pay 1000.”
To design the personas we used Cooper's methodology. This types of personas are based on heavily in user behaviour. Demographics are only used if it's really necessary.
To turn user behaviour into personas, we make a behaviour map, based on users answers. These are the steps:
Two personas emerged in the mapping, one primary and a secondary. This means that the solution we design for the former also benefits the latter. Maybe not completely, but with some customisations we could have a great product for both types of user.
Research took us to three main attributes we believed our product should have. These attributes would guide throw all design phase. Well, not only the design phase, I think they turned to be the teams mission in a sense.
Users can mainly do 3 things in a human-computer interaction: consume information, manipulate information and navigate throw information.
So here's what I did.
First we tried experiment 3 interfaces: A simulator, a chat bot and a list of proposals.
Simulator won. Chat bot would be interesting for the emotional aspect, but the user interaction could feel to linear and lack the control we wanted people to have. For the list of proposals we found out that the amount of information was kind of overwhelming.
After deciding for the simulator, we iterated a few times and went to user testing. The winner was the one with the stepper. The minimalist design made it the most responsive.
I focused in the hiring part to avoid a long article, but here's a grasp of the whole product.
the benchmarking, hitting 10× during roll-outs
was the NPS for customers that got to the simulator
“I want to thank you all for this dream made true. Yeah, we bought the land-lot! I just wanna say thank you, thank you, thank you.”