As Principal Product Designer designing a new product, I worked with a UX Researcher to interview and observe several users as they walked us through their data operations workflow.
The insight we uncovered is that a team of typically 5 or fewer developers and 1 manager have multiple data jobs in development and production. After design and deployment of the data jobs, the team is also responsible for monitoring the data flow on a day-to-day basis.
The overview dashboard gives the data integration team a quick overview, or what I have termed the Monday morning check of data integration jobs.
I spoke with several dozens of customers prior to the design of this application, but the validation never stops and the best reaction after the product launch was this:
What are the “right kinds of metrics”?
Rows Processed and Number of Jobs are the two key metrics. Shown above are miniature visualizations of the data to be used on the overview dashboard. Hovering over the columns provide additional information. The Number of Jobs chart displays an additional dimension of data by breaking down the count by ‘job status’, so a stacked column chart is the most appropriate visualization to show the states Success, Warning, and Failed.
Each of the key metrics can be correlated with supporting attributes such as ‘projects' or ‘tags’. The line charts above let users visualize and compare the data over time.
While interviewing and observing customers, I observed a variety of solutions that developers have pieced together to support their routine workflows.
This contextual observation generated numerous ideas for data visualization and data reporting features, and also inspired new product strategies such as an Informatica Slack app for monitoring.
Utilizing Slack gives teams an impactful way to view Informatica operations data in a collaborative environment that they use daily anyway.
Big Numbers, 1st Glance
A key insight I discovered from observing users is that they want the first glance to contain large actionable numbers, or “fast facts”, rather than common visualizations such as pie charts and donut charts.
Analysis Views, 2nd Glance
Visualizations are the second glance for a more detailed analysis of the data.
Earlier mockups use company logos in the y-axis labels. This was a popular idea among product marketers. However, the design proposition wasn’t very scalable considering Informatica supports several hundreds of connector types, and custom connections wouldn’t come with an associated logo image.