Home

Research Methods

Depending on the project, I performed various types of research for The Benefit Bank. The following are some of the methodologies that came into play:

• online ethnography
• contextual inquiry
• card sorting
• diary study
• customer interviews
• surveys
• competitive research

Insights

Research helped to create some general heuristics for creating designs in The Benefit Bank. The data could be distilled down to form the following insights:

Architecture

Concepts were refined using various architecture methods. Mappings were used to better visualize processes, how information flows, and to create communication artifacts that could be shared with various stakeholders.

Steps Document

There are several types of "portal docs" in The Benefit Bank, including "steps", "checklists", "fact sheets", and more. There are hundreds of these documents in total, many with varying formats and inconsistencies. So, the challenge of unifying these varied documents was the primary focus. Also, regarding the contents of the document, it was a challenge to figure out the hierarchy. I worked with the domain experts to question if any information could be removed from the often dense text or if anything in particular was important enough to be called out. I also worked out a new way to categorize the numbered steps into groupings that would better match the user journey.

New Path Forward

The new design came together with slight variations for both various device resolutions and also testing done to make sure they would print well. These documents were often printed to either get people prepared for using The Benefit Bank or as a take away for after they were finished.

Tax Survey

Each tax season, I typically work on adding in new features and also sussing out users' pain points. In order to aid in doing this, I developed a survey in collaboration with a client, and have had it collecting data for 9 tax seasons. It's been an invaluable source of data to get a sense of how the tax service is doing, investigate issues, and make note of new potential features. Typically, between 10-15% of filers will fill out the survey, so there's a lot of takeaways.

For example, occasional I have discovered edge cases that were not accounted for during the complex tax law and research planning phases. One respondent said that she, "failed to file federal because 2 different people have W2's". It took some effort to figure out that this was a woman that could not file her taxes with her adult child being included on her tax return. The child would need to file a separate tax return. While this is a rare use case, it revealed that our help content on the subject came much too late in the process. The information was moved to the beginning of the process instead, to make the process of filing more efficient for people like this who may need to make adjustments.

E-file Status

Through the tax survey, consulting with customer service reps, and other stakeholders working on taxes, I also determined that there was an ongoing issue where users thought they had filed their taxes, but in fact they had not done so. So, for 2018 taxes, I worked on making the users' tax filing progress more clear. I recommended many changes to text throughout the tax filing process, including at times, reconsidering the flow of pages.

One interesting addition was adding the idea of taxes being "started" vs. "not started". The idea seems simple, but was not ever connected to the idea of an e-file status. E-file status means that once taxes are submitted, they can be in the state of "processing", "sent", "accepted", "rejected", and a couple of more rare statuses. However, if a user was in the process of filing and had not yet sent the e-file, it was not immediately clear that the user had begun the filing process. I implemented a more pared down version of the e-file status, where the idea of "starting" the filing was added right next to the link that takes one into taxes.

Login Recovery

Users not being able to access their accounts is a common issue for many applications out there. For TBB, this can at times be exacerbated by users that do their taxes and only log in about once a year. The customer service representatives get inundated with a large amount of calls during tax season about log in information, so any changes that could lighten this burden were welcome.

I first did a basic cognitive walkthrough on the site, going through the steps of users both forgetting their password and/ or forgetting their username. I made a couple of changes right away, including adding another set of links to the “forgot login” and “forgot password” tools. Another research step that I took was actually helping out with customer service for an hour each day for a few weeks.

Problems Solved

Not surprisingly, I was able to work out more insights by getting into the trenches. I made note of several ways to make changes in the app, that would make the customer service representatives' job more efficient. For example, I advised that a change be made to the search by SSN for users. The search would work with only 4 digits, but these 4 digits did not only look at the last 4 digits of the SSN. This meant that sometimes multiple search results would appear if the 4 digits matched any part of the SSN.

I also made note of changes we could be making that would give users more of an ability to access their account by themselves. For example, I noticed that in one of TBB's service models it was possible to answer security questions to get into an account, while in another service model, it was not possible. Also, at times there was text that simply gave the customer service contact information instead of linking to the self-service login and password recovery tools.