FOCUS Brands BI Intelligent Analyst
Design Scope: Dashboard redesign
Tools: Sketch, Illustrator, InVision
My role: UX designer, UX researcher
Project Type: School project with the client
Currently, the BI tool at FOCUS Brands has 20 dashboards which is used by 6 different departments. As part of reporting, these dashboards can help employees view and dig into data relevant to their areas of work for insights and trends. But as all new implementations do, FOCUS Brands is currently challenged by the task of increasing the adoption of Power BI across the corporate level. Consequently, FOCUS Brands wanted our team to help them research and come up with possible solutions to make the dashboards more effective for their employees.
The Wizard Assistant is an artificially-intelligent digital assistant that supports all levels of business users in FOCUS Brands to solve any frustrations they may encounter while using the Power BI dashboard.
Raise requests and keep in touch with BI team
The function is to reduce the back and forth in-person communications with the BI team. Once the user ask to request a new report, they will fill out a form in the chat, which will be submitted directly to the BI team.
When the user types in descriptions into the chat, the system can 1) directly lead the user to the certain places in the reports and 2) find relevant reports and data for the user
After receiving this design challenge from our client, the group went through 4 phases to identify the problem space and to find out design opportunities.
What was the BI dashboard?
With little related background knowledge, the group then studied 8 extant literature on business intelligence and data visualization.
Who were the users and stakeholders?
Based on our investigations and the orientation given by Mr. Kuester, we came up with the following user categories
During the phase II, a series of research methods were deployed to gain more insights on the contexts and needs of the employees at FOCUS Brands.
How did our users carry out their work practice with current dashboards?
Although we had an inkling of how these dashboards were used from the words of Mr. Kuester and through initial background research, we had never seen how the users actually interacted with the dashboards. As such, we decided to conduct an on-site contextual inquiry, since putting the user in the expert role helped us learn more about how a typical user went about using the dashboards. We also had a chance to perform semi-structured interviews with 3 business users to obtain detailed insights and attitudes on Power BI.
How did we analyze the data and what insights did we get?
After conducting the onsite research activities, the group decided to utilize affinity mapping to find patterns to the issues, requirements and get to the mental model of the FOCUS Brands employees.
Key takeaways from the affinity mapping
Throughout the research process, the group were able to figure out the root causes behind the low adoption rate of the current dashboards and prioritized these problems based off of their severity.
Who were important?
Another big set of findings that we generated from the affinity mapping was that we realized there was a huge expertise and knowledge gaps among different business users due to their background and work. In this way we created 3 types of user groups to help us find their needs as well as our design specs.
After visualizing the personas of each kind of users, the group figured out the second user group made up the majority of the employees, addressing their issues would directly impact the adoption rate. Meanwhile, by solving the target user group’s issues, the resulting benefits will trickle down to the secondary user groups.
How can we address the user pain points?
Moving forward, the group brainstormed possible solutions with research findings and user groups in mind. Later, we organized and converged all the ideas we came up with and came up with 4 concepts.
How did we narrow down our concepts to the Wizard?
Following the user centered design principle, the group decided to give our users the right to choose their favorite concept. However, the responses were varied according to their own preferences. Since we were more concerned with the non-tech savvy users, we wanted to make sure that eliminate the inertia involved with learning something new. A lot of users already had some existing alternatives and relying on human help. So we thought why not redesign that human help and still be more efficient.
How did we develop the wizard concept?
First, building on design implications, we came up with four important features to help with.
Then, since we were designing the artificial intelligent assistant, we should start with the personality of the assistant. How would it behave? How it would converse? What will be attitude towards different types of users. How can we make people trust the assistant?
Then, we started prototyping…
The group started designing with the low-fidelity prototype according to the features we defined earlier.
In order to test the usability and feasibility of our idea as well as to get feedback for future iterations, the group decided to conduct a heuristic evaluation and a usability testing. We invited 3 experts and recruited 6 users/stakeholders to help us test the prototype. The overall feedback was positive; however, some of them were still concerned with the following 3 issues. Later, the group made some improvements according to these issues.
Add more features to help with the querying process
The group add quick access options that can provide possible actions for users. Meanwhile, we add an auto-filling mechanism. Once the user starts typing, the Wizard should be able to provide suggestions and/or relevant.
Change the entrance of “Raise a request“
Some users were concerned with the button of “Raise a request“ because they have no expectation of the form. Thus, the group decided to disclose the image of the form.
Add the preview for the files
During the user testings, some users pointed out they would like to have a quick preview of the file to see if this is what they want. As a result, the group add more actions to meet as many users needs as possible.
The intelligent analyst is a designed system that wasn’t simply tweaking the UI of Power BI. Consequently, the process of designing the final prototype had a layered approach. From the exterior aesthetics to functional features to contextual user case scenarios.
Things we learned
Working with our client FOCUS Brands was amazing. Our team learned a lot from UX research methods to team project planning during the whole process. Here are the three lessons I learned:
Working with real clients means we need to be flexible all the time
Our project planning did not account for the potential delays that could emanate from client’s side. We planned a lot of user studies such as diary studies which were later refused by the client due to their tight schedule. We should have anticipated such scenarios and prepared ourselves with rescue plans.
Iterations bring powerful results
The whole process was an iteration, and we should redeem each phase interlocked with each other. For example, we failed to gather quantitive data while doing the user research, but we later realized that this was iteration one of the research endeavors and that more iterations would be happening to fill the gaps from this stage.
Let users help you make design decisions
Always put users in the center of the decision making process. Since we were not the end users, what we did during the whole process was to keep learning from the business users. This master and apprentice relationship guided us through every decision making steps.