Working in a time crunch, digital marketing analysts want to see their clients' key performance indicators (KPI's) quickly and accurately. Visualizing basic information can be frustrating and waste precious screen space for analysts. But more advanced visualization helps the analyst tell a story about how the marketing campaign is progressing, and make predictions about its success.
At Centro, I interviewed multiple analysts and worked with them to identify questions they'd like answered about their campaigns, and to create examples of their ideal visualizations. Included here are some sketches from this work - from user feedback and investigation, to early sketches, to example mockups.
Talking to users regarding analytics brought to light the differences in potential customers. Analysts are comfortable with numbers and already know what to look for in the data. Advanced charts with dual axes or tiled information help these experts gain deeper understanding. But not every company uses analysts - indeed, there is often one person in charge of campaign planning, execution, and analysis for all campaigns. These users are pressed for time and want quick access to basic metrics. Information should be quick and self-explanatory, and any campaign success or failure should be highlighted.
Snapshot of the Analytics page without visuals.
Visual shows the basic health of the campaign: what revenue will remain unspent based on the current performance.
The above bar chart helps advanced analysts and novices alike get an overview of when certain publishers should be running.
As opposed to a perfect schedule as shown in the bar chart, these stacked sparklines represent delivery from each publisher.
Basic delivery over time, visualized as an area, can be useful for seeing simple trends in the campaign, like spikes or slowdowns in impressions delivered.
Beyond simple delivery, analysts want to see how their specific KPI is performing over time - whether it is CPI, likes, total conversions, etc.
Dual-axes graphs like these are a favorite of advanced analysts, because they tell a fuller story than only a single metric. Advanced analysts don't simply report delivery; they understand performance and how it is affected by that delivery over time.
Centro Platform now has an overview of the campaign, using donut visuals to convey basic information quickly, although users can drill into the analytics to get more detailed data.
Statistical analysis is ready to be the next leap in business intelligence. Existing statistical products are very complicated - focused more on test names than on utility, there is a gulf between people who understand statistical analysis and those who don't. This is why we created Analytic Catalyst at IBM SPSS in 2013. I was the lead UX/UI designer, and our goal was to design a product where business users can make use of statistical analysis without being trained in it.
In 2014, Analytic Catalyst was brought into Watson Analytics as the predictive portion of the product. I was head UX Architect of the predictive portion, tasked with working with a BI group to meld Analytic Catalyst into the Watson Analytics product. Together we made additional enhancements, such as a cleaner look and feel, reduced information on a single screen, and a cohesive Spiral Visualization that combined model-based analysis with correlational analysis.
Below are the steps I took to arrive at a product for Analytic Catalyst, and then as the predictive portion of Watson Analytics.
Layout and workflow created by me. Final visual design created by a visual design partner.
Decision trees visually explain how data is separated into sections in order to better predict an outcome. Users are able to select a section of the tree and drill into the outcomes that exist there. While advanced business analysts liked this description, users who were untrained in statistical analysis wanted to see something more clear-cut: 'tell me which clients are most likely to leave.'
A later redesign of a decision tree, after talking with users about how they want to visualize their data: 'Show me answers first, then explain why, and let me play around to find my own insights.'
My parents have several different pills to take. Eyesight is poor, and keeping track of changing medications is tricky.
Current medication apps focus on reminders and refills, and don’t work well if I’m not the one taking the medication.
Shown here are the steps I've taken to create a new medication app that allows the elderly (and their caregivers) to add & update medication easily for reference when taking pills, filling doses, or communicating to doctors.
View the prototype on Marvel or the prototype on InVision.
