Introduction: Context
Focus Brands is a food franchisor company owning 7 different well-known restaurant and snack brands such as Auntie Anne's, Cinnabon, and Moe's Southwest Grill. At the massive scale with over 5,000 stores in more than 50 countries around the world, data is an essential tool for Focus Brands to run their business.
Recognizing this necessity, the company is currently trying to promote the use of data to its employees, hoping that it will help them make well-informed business decisions and pushing the business forward. Focus Brands established its own business intelligence (BI) team and a BI tool based on Microsoft Power BI a few years ago.
Note: Business Intelligence (BI) tools are applications that help users analyze data by processing and displaying them into visualizations, helping employees make well-informed decisions based on actual data.
Overview
Problem
According to Focus Brands’ Business Intelligence team, the tool is still suffering from low adoption rate. This means that the company's investments in the BI system could be wasted without achieving the goal of more data-informed decisions in the company.
Approach & Goals
We aim to solve the adoption rate issue by redesigning the tool to be more useful and usable. With a user-centered approach, we discovered multiple pain points that hinders the growth of the adoption rate. Our design aimed to
- make the data displayed in the tool more relevant and easier to digest
- facilitate collaboration within the team and externally
- help user understand the data source, thus building trust in the tool.
Final Design
Reducing cognitive load, facilitating communication, and building trust in the BI tool to help Focus Brands promote a data-informed business culture.
Customizable Dashboard
"I look at these data a lot. Let's move it together in one place"
- First-time users will be asked a set of questions about their work and preferences to help them get on board and customize their home dashboard.
- Quick tutorial to kickstart users with the tool.
- Customizable dashboard in which users can pin charts from other places in the BI tool to their home page.
- Showing only relevant data in their home page to avoid overwhelming users.
Search & Navigation
"I need the in-store sales data for this brand but it's not in my home dashboard yet... Let's do a quick search."
- Users can search the whole BI system with keywords from the search bar on top of every page.
- Making it easier to navigate to a set of data in the BI tool.
Goal Tracker
"It's been about half way through the year. How is my team performing towards the store participation rate goal?"
- This tool provides an easy way for users to keep track of their goals.
- It shows how the goals progresses throughout the timeline of the goal.
- Potentially related data are also suggested automatically based on correlations.
- Helps make the data more focused towards users' goals.
Comparison View
"These data seem correlated. Let's take a look at it side-by-side"
- Side-by-side comparison view for users to look at multiple set of data simulteneously.
- Lets users create focused, uncluttered views for the data relevant to their needs.
Data Sharing
"Hey, I found some interesting trend in delivery time for this week. What do you think?"
- Provide a seamless way for users to communicate with other people about a piece of data in the BI tool.
- Users can also annotate the charts before sending them to other people via emails, the internal communication tool they already use.
- Facilitating the communication about data between employees.
Export for Presentation
"It's that time when I have to make presentations again. Let's do it quick!"
- Users can copy chart images and paste them to presentations directly.
- They can also select an option to annotate the charts before exporting them as an image to be used in other apps.
- Ease the tedious user flow of using the data from the BI tool to make a presentation.
Support Request
"Some data points on this chart doesn't seem right... Could you help take a look at it?"
- Providing a way for users to communicate with and express their needs to the BI team.
- Also benefit the BI team in listening to their users and continuously improving the tool.
Metadata
"Hmm... When was this catering data last updated? How did they calculate this number? Let me check..."
- Users can see information about data including when it was last updated and where it came from.
- Building trust by helping users understand the reliability and freshness of the data as well as understand possible causes of data inconsistencies.
- Also shows calculation formula for numerical metrics in the BI tool to help users better trace and understand the data.
Process Overview
1. Empathize
- User Interviews
- Steakholder Interviews
- Contextual Inquiry
- Surveys
- Affinity Mapping
- Personas and Empathy Maps
2. Ideate
- Design Goals Definition
- Brainstorming & Ideation
- Sketching
- Wireframing
- Feedback Gathering
3. Prototype
- Sketching
- Wireframing
- Feedback Gathering
- Interactive Prototype
4. Evaluate
- Usability Testing
- Heuristic Evaluation
- Design Recommendations
Empathize: Understanding Users & Context
We started the project with exploratory user research to dive into what are the needs and pain points of our user group related to the BI tool.
Who are the Users and Stakeholders?
Users
Focus Brands Employees
- Works for a specific brand or in a centralized teams shared across brands.
- Have different domain of work, e.g., restaurant operations, marketing, finance, international
Stakeholders
Focus Brands BI Team
- Centralized team shared across different brands.
- Responsible for development of the BI tool
- Process data from source and display them in the BI tool for other employees to see
What We Wanted to Know
The research goals we addressed included:
- To understand the nature of the work of the employees.
- To understand how business data fits into employees’ workflows.
- To understand the current use of the BI tool.
- To understand the BI tool itself.
- To understand the work environment of employees at Focus Brands.
How We Collected Data
From the goals, we then planned our research in the planning table shown below. We started by listing questions and information needs for each goal and then pick an appropriate UX research method for each question.
Below are the methods we actually used:
User Interviews
2 remote + 2 on-site semi-structured interviews, 40-60 minutes each with Focus Brands employees
To learn about their day-to-day work, business information needs and ways of acquiring them, how the data is being used, and channels of communication within the company and to external partners.
Stakeholder Interviews
60-minute Remote, semi-structured, group interview with the Business Intelligence team.
To understand the needs, plans, and challenges of the BI team as well as how the data is put in the BI tool and the rationale behind the design decisions made in the current system.
Contextual Inquiry
1-hour on-site contexture inquiry session with a Focus Brands employee.
To understand the workflow and the tools they use in their day-to-day work as well as the work environment of the user.
Surveys
8-minute survey sent to the broader set of employees at Focus Brands
Unfortunately, we only received 4 responses
The goal was to reach more participants and validate the qualitative research with quantitative results and learn about the employees needs and pain points with the existing BI tool.
How We Processed the Data
We used affinity mapping to analyze the qualitative data we collected from the research methods we conducted.
Combining that with a little quantitative data we collected, we distilled a list of user needs and pain points as shown below.
Personas, empathy maps and storyboards also helped us put these needs and pain points into context to better empathize with our user group. We created three personas reflecting different user segments, e.g., managerial/operational employees, frquent/infrequent users.
Insights
User Needs
Opposing to our initial thoughts, the user needs we discovered suggested that the use BI tool is not just for "looking at summary of data" and making decisions directly from them. Instead, there are also aspects such as measuring performance, communication, and creating presentations and reports, that are heavily involved in the use of the tool.
The user needs that we found included:
1. Understanding the Business Situation
Employees use the BI tool to keep track of the performance of their campaigns and see whether they have reached their goals/KPI.
2. Business Decisions
The data in the BI tool have helped employees make decisions to grow the business, e.g., deciding on the menu line-up for catering.
3. Communication
Doing business is a team effort; they need to communicate to progress towards their goals. These communications happens both within the company and with external partners.
4. Presentations and Reports
Communications also take place in more formal forms such as presentations and reports. These are to show the data and communicate ideas to higher-ups and people they work less closely with.
Pain Points
We found the problems that could annoys and/or hinder the users from completing their tasks:
1. Confusing Navigation Schemes
Participants repeatedly mentioned about how they do not completely understand how to navigate around the BI tool.
2. Overwhelming data
With the amount of data presented in the BI tool, some data seem irrelevant to the users and, in turn, the data feels overwhelming.
3. Presenting data
Compiling data for a presentation is a tedious and time-consuming task.
4. Sharing data
Getting data from the BI tool and transforming them to a more shareable format is also a taxing task for the users.
5. Data Reliability and Availability
There are inconsistencies in the data presented in the BI tool; some data points displayed are not consistent across different views, or some might conflict with other sources or users’ prior knowledge. This make some users feel skeptical about the credibility of the tool.
6. Learning the BI Tool
Participants often mentioned that it "took a while" for them to learn how to use the tool.
This journey map further illustrates the steps a user's would have to take and pain points that could occur when they are using the existing system.
Defining the Design Goals
We compiled the findings and defined three main design goals to help our team focus our design efforts on. These design goals include:
1. Reduce Cognitive Load
Help make data more relevant and less overwhelming
2. Facilitate Communication
Make it easier to share data with others and provide channels to contact BI team
3. Build Trust
Help users understand source of data
Design Ideation
Based on the user needs, the pain points, and the design goals we have defined, we brainstormed design ideas for possible features to address those goals. Some of the feature ideas that seem viable were then selected and grouped together, organizing them into three design alternatives to be prototyped.
Iterative Sketches & Wireframing
Sketches and wireframes are great tools for communicating our design ideas to other people. These tools helped us conduct effective feedback sessions with our user group and evolve our design before investing the time to create more refined prototypes. Two iterations of wireframing and feedback gathering were conducted with increasing fidelity of the wireframes in each round.
Feedback Gathering
For each iteration, we conducted feedback sessions to understand users’ thoughts on the design ideas we came up with. The feedback were then used to improve and evolve the design and create a more refined prototype for testing.
1st Iteration: Low-fidelity Sketches
The 1st iteration was a divergent design ideation effort. For each design goal, we tested three design alternatives addressing these goals and collect feedback about them.
Goal 1: Reduce Cognitive Load
- Alternative 1
- Goal-based BI: Helping users focus on their business goal by letting them set and track their performances against the goals. The tool also suggests data relevant to each goal.
- Search Bar: Keyword-based search functionality to let users easily navigate to a desired data easily.
- Alternative 2 - Onboarding Quiz: Using a quick questionnaire to understand users' data needs and help pick only relevant data to display on their personalized home dashboards.
- Alternative 3 - Onboarding Tutorial: According to user research, users do not understand the navigation scheme of the BI tool. This design idea aims to help solve that by displaying a short overlay tutorial to first-time users.
Goal 2.1: Facilitating Communication - to Colleagues
- Alternative 1 - Sharing, Exporting, & Annotation: Make it easier to share a chart on presentations, reports, or via emails by providing an option to do that in which each chart. Users can also annotate charts before exporting them to help them communicate better with the recipients.
- Alternative 2 - Clipboards: Allowing users to "pin" multiple charts to a single "clipboard" before exporting/sharing them with their team.
- Alternative 3 - Drag & Drop Sharing: Users can drag and drop a chart to an area on the navigation bar associated with a team to share it with that team.
Goal 2.2: Facilitating Communication - to the BI Team/Administrators
- Alternative 1 - Feedback form: Make it easier to share a chart. Lets users send messages directly to the BI team through a simple feedback/contact form.
- Alternative 2 - "Can't find what you're looking for?": Provide a centralized way for users to request for missing/unavailable data by clicking on a button on the navigation bar.
- Alternative 3 - Data Request Form: A data request form which allows user to specify which data in what preferred type of visualization is needed.
Goal 3: Building Trust
- Alternative 1 - Data Source Information Footnote: Shows metadata including source and last update time at the bottom of each chart.
- Alternative 2 - Data Source Hover: Shows a similar set of metadata to the users when they hovers the cursor on a data point.
- Alternative 3 - Anomaly Notification: Notifies user when anomaly in data trends occur.
What did the users say?
There were a total of 10 participants joining our feedback sessions. Some points mentioned by the users during this round of feedback session included:
Goal 1: Reducing Cognitive Load
- The goal tracker is a valuable tool for their workflow. Users mentioned that the goals usually involve more than one person in the team.
- Some questions in the onboarding questionnaire were not neccessary since it already depends on the data permissions they have. Users expect an option to add charts to home dashboard after the quiz.
- Some users mentioned that they would skip the tutorial while some novice users found the tutorial helpful.
Goal 2.1: Facilitating Communication to Colleages
- The users found the sharing/exporting/annotating feature useful.
- They also mentioned that the clipboard feature would be useful and mentioned that they want a side-by-side comparison view for different charts.
- Some participants were concerned about sending data to other people who do not have the permission to view such data.
Goal 2.2: Facilitating Communication to the BI Team
- Users are concerned that their feedback will be missed. However, the BI team said that they welcome more feedback from the users.
- Some users are concerned about conflicting requests made to the BI team for data visualization types.
Goal 3: Building Trust
- They find displaying the metadata along with the charts helpful for evaluating the reliability of the data. Some also mentioned that they would want to know how the data points were calculated.
- Users prefer to see metadata up front without hovering on the data points.
- A clear anomaly indentification technique must be defined for the anomaly notification to work properly. Participants are concerned that they will either miss important updates or were flooded with too many notifications.
Prototyping
A prototype is a crucial tool to help us evaluate our design ideas in the next stages. Adobe XD was used to create an medium-fidelity interactive prototype of this design.
2nd Iteration: Medium-Fidelity Wireframes
Based on the feedback from the 1st iteration, we merged our design alternatives, converging into one set of medium-fidelity wireframes which were then used in another round of feedback gathering.
What did we changed?
Here are some examples of what were changed in this iteration:
- Most feature concepts were merged together into this one design and were implemented with higher fidelity to show more details of how they will work.
- Some feature concepts were altered based on the feedback in terms of user interface layout.
- Drag & Drop: Allow other features to be triggered with this interaction style, e.g., drag a chart to the "Support" tab to send it to the BI team.
- Pin to Home Dashboard: Let users pin/unpin charts to/from home dashboard after the onboarding quiz.
- Goal Tracker: Allow users to set goals for the others, e.g., a manager setting a goal for their team.
- Notifications x Goal Tracker: Use notifications to notify users about goals instead of using it for data anomaly.
- Support: Combined all support features together and added a status for a case, e.g. "pending", "acknowledged", "solved", to remind users in case they need to follow-up with the BI team.
- Visualization Type Request: Instead of having users request the type of preferred visualization technique for a set of data directly to the BI team, provide them with toggle buttons to let them do that on-demand instead.
What did the users say?
Here are some example of the results from feedback session/informal testing for this iteration:
- Most features were useful and easy to use.
- Participants did not know that they can use drag-and-drop to perform actions, instead, they used the more visible context menu icon as an entry points to chart-based actions.
- As a result, this feature was removed in the final design to avoid confusion.
- The name of the "Clipboard" feature confuses the users. Once understanding how the feature works, the users find this feature useful.
- The menu label for navigating to this feature was changed to "compare data" instead.
- The search bar, which was designed as a gray blank underline on the black background of the top bar, was not noticeable enough.
- Change the visual design of the search bar so that it is a white box against black background.
- In goal tracker, the layout in which the goals were presented—laid horizontally as tabs—cannot effectively handle the number of goals an employee will have simultaneously.
- Change the layout to accommodate more goals by stacking them vertically in the sidebar instead of horizontally as tabs on the top of the page.
Revised Interaction Prototype
The final design was updated from the 2nd iteration wireframes. Changes were made based on the user feedback that we gathered.
Overall, similar to the existing tool that Focus Brands is using, this new design is a web-based tool which displays business data curated by Focus Brands' BI team. It provides a more focused workspace for the users to see data relevant to their work to streamline their workflow. This design also aims to pursue the other goals of reducing cognitive load, facilitating communication, and building trust in the BI tool.
Visual Design
As the layout of the design has been laid out, we began looking at visual design of the system to create an aestheticly pleasing experience to the user. These are some design directions we explored.
Evaluation & Outcome
With the prototype, we put our design to test by conducting both user-based and expert-based evaluation methods.
Moderated Usability Testing
We conducted 3 on-site and 3 remote, 60-minute usability testing sessions with 8 tasks. Participants were asked to rate each task in terms of ease of use and rate the realated features in terms of usefulness.
System Usability Scale (SUS) questionnaire is used to get an overall benchmark for usability of the system.
Heuristic Evaluation
We also conducted 3 expert feedack sessions based on heuristic evaluation techniques to help us spot design flaws we might have overlooked. Neilsen-Norman's 10 Usability Heuristics were used to guide the session.
For this class project, the experts we consulted consisted of 2nd years MS-HCI students who have experience in usability and have experience in UX from jobs and internships.
Testing Results
Overview of Results
The quantitative data we collected from usability testing included Likert scale ratings of ease of use and usefulness of each feature we tested, and SUS scores for to assess the design as a whole. The results showed the users' overall satisfaction with the proposed features in terms of ease of use and usefulness.
Insights from the Feedback
Qualitative data we collected helped us identify the problems in detail. After collecting them, the data were grouped based on related features and ranked by severity of the problem. Some issues we found included:
1. Explore Data
Issues- When users navigate to a chart not existing in their home dashboard, most of them did not think this menu will show them the data they want.
- This wording problem is also the only issue which received highest severity rating in expert evaluation sessions.
- Change the wording from “Explore data” to a more straightforward term such as “All data”.
2. Comparison View
Issues- Some test participants did not understand how the feature works.
- Having a “Compare Data” menu in the navigation bar tends to mislead the users and confuse them.
- Some experts also mentioned that the functionality of this feature could be confused with home dashboard since both work by having user adding a chart into some set.
- Remove “Compare Data” from the navigation bar.
- Only let users add charts to comparisons through the context menu (“Compare…”). By clicking on the menu, a bottom bar will pop up. This bottom bar contains the charts existing in the current comparison. Users can save the comparison to their home dashboard to view later.
3. Freedom of Navigation
Issues- Both participants and experts mentioned the lack of back button and inconsistent appearance of breadcrumbs in the design.
- This made it more difficult to navigate back to the previous page if a user made a mistake.
- Add back buttons and/or breadcrumbs to every page.
4. Goal Tracker
Issues- All participants found this feature very useful, however, some participants mentioned that they want to prioritize the goals they set as not every goals are equally important.
- Allow users to drag and drop tasks in the list to rearrange it.
- Provide options to set the priority of each task.
5. Data Sharing
Issues- Participants mentioned that they might need to format the email before sending it and they expect to do it in their email application, i.e., Outlook
- Change the label of the existing button from “Share via email” to “Send email now” and add an “open in Outlook” button to give them the ability to further format the email.
Next Steps
- Fix the design issues:
- We have identified the issues in our design through different methods of evaluation, the obvious next step is to fix the issues according to the findings for better user experience.
- Details & Prototype improvements:
- At this stage, most design efforts were spent on the conceptual side of the features and not as much in terms of exact details.
- In future iterations, we should look more into detail, e.g., what exactly should be asked in the customization questionnaire, to refine our design.
Key Takeaways
- I learned about UX Process and evidence-based design by conducting research and designing a product based on the findings.
- I learned how user research could help us focus our design efforts to the problems that actually matter to our users.
- I learned to work in a diverse team with team members having different backgrounds.
- I learned about the constraints of working in a real business environment.
Acknowledgements
We would like to gratefully thank...
- John Kuester, the Director of Business Intelligence and Analytics at Focus Brands, for facilitating this project and providing us the resources we needed.
- Focus Brands employees who participated in our research for providing us with valuable insights for our design.
- Huaiwei Sun, Ishaani Mittal, and Ruturaj Eksambekar, for their help in the heuristic evaluation sessions.