The Problem

When I started working at this startup, there was very little research and no guidelines on what to research on. The designers and me exclusively worked on problems they could easily identify, which, as I later found out, left out some of the key problems that kept people from signing up.

The Process

Since there was little awareness for research, usability testing and so on, I took it into my own hands to set up processes and look into ways to identify the biggest problems.

I worked closely with the support and the business intelligence department and established communication. I also gathered all the info from Marketing I could get. Together, we figured out a key problem with one of the target groups of the app and I went on to investigate. We also figured out a key strength of the platform no one saw before, which only got obvious once I collected the data in a clear way and created visualisations that were easy to digest.

I did some visualisations in d3.js, Tableau and with python to communicate the problems and possible solutions to stakeholders.


Personas can be misused quite easily and I worked especially hard on mine.
The preconditions for personas weren't good: All the data I had was mostly from the database and not from users directly and there were no resources for proper research

Why did I even wanted to do Personas then?

In my research I figured out that the people using our product were widely different from the expectations of my co workers.

With my personas, I mostly had the goal of showing stakeholders and fellow co workers that our users are more diverse that what people in general assumed.

I found that when only showing the numbers that prove that, people only concentrated on the numbers which were aligned with their expectations. The personas, together with the numbers, were supposed to show that all of these different user groups were equally important.

I found users on twitter, facebook and reddit. This approach was of course just a starting point, since I assume many of our users are not active on any social media platform, but I understand my personas as something that evolves constantly anyway. I talked to them, read their tweets and tried to figure out their problems. Not only specific problems with our platform, but general problems in their life that we could solve.

I used every opportunity I had to learn about potential users. How they view investing in general, how they see crypto and what their goals are. I talked to everyone who wanted to discuss it and gained lots of valuable insights, also many of which weren't crypto specific – but I knew that our product wasn't aimed on only solving the needs of people who wanted to use crypto, but also people who wanted to take their financial future into their own hands.

Usability Testing

Since there were no clear guidelines or any examples of usability testing before, I set up a document to make sure that all my research on the topic in the future can be still used.

This is what I came up with and what my team also helped me refine:

Step 1
Short introduction on what we do and why we test (without being too specific to influence their opinion).
Make it clear, that they are not judged and this is not a test and they are not being tested, WE are being tested. Have water and everything ready. Introduce them to the test format, ask if we’re allowed to film or to record the test. Set everything up.
If they have questions, they should ask after the test. Ask them to think out loud.

Step 2
Start a conversation about us and how they intend to use the platform, let them set their own goals. Settle on a task together. Find out the context of when/how/why they are using the product.

Watch closely and when necessary ask questions to dig deeper into the thought process of the user. Encourage them to talk about their feelings/thoughts (especially important while they're going through the process..). Avoid leading questions such as “Did you find that difficult?” to not influence their opinion or make them feel insecure about expressing their opinion when their feelings contradict the leading questions tone.
If they start steering the conversation towards stuff like “I guess other people would struggle with this”, firmly lead them back to them sharing only their experiences.
Never judge, but also never positively or negatively reaffirm their feelings. Stay neutral throughout the process.

Step 3
Answer every question they have, general discussion about the specific problems, what their thoughts on the topic are, what their experiences are, etc…

Possible Problems:
When testing, always be aware that this is a situation (most of the times) different from their usual preconditions. They are fully focused (which isn’t happening often) on the task and the test spaces are often sterile and they are being watched.

Always set a clear goal for every testing session!

Tech spec:
Quicktime for Screenrecording (Mac)
iPhone for recording audio

This worked okay-ish, but I had a hard time getting all the points across I wrote down in the document.

My mistakes mostly happened because I got nervous when when users wanted to go quickly through the test, which lead me to skip through my document and forgetting some of the most important points.

As a solution, I started writing checklists before every testing session, to make sure I wouldn't forget to mention anything and had a clear timeline through that as well. So far, my process got better, but I'm still looking forward to refining it even more.

The Result

click here for a higher resolution of the personas

There is no final result, because this is an ongoing process.

My personas aren't in a fancy template (but in google docs) because I wanted to keep them very open for changes and accessible to everyone inside the company. I also refrained from using too many details, so people wouldn't be drawn to stereotyping as much. I also created a gender neutral version of the personas, but found that these worked better for my goal: Creating awareness that not all our users are young able bodied white men who trade crypto regularly just to get rich.

These personas still need to be double checked with more data in the future, but are a starting point. For many of my tasks, jobs-to-be-done are more successful and useful than these personas, but for the sole goal of giving everyone inside the company archetypes (and their goals) to work with and understand their needs and pains, this proved to be successful.

I created them by gathering general research about investing, usability testing sessions, discovering our users on social media, by reviewing support tickets and by reviewing data from out data base.