It’s a common misconception that if you’re harnessing data, you’re giving users information. But data alone is not particularly useful. It’s only when we add context that it can be translated into something that informs our business decisions.
So how do we make this transformation for our users? The answer is in analytics dashboards and visualizations that go far beyond traditional graphs and bar charts. When we present users with data they can easily access and understand in the context of their workflows, we empower them to glean actionable insights.
In the last decade, the variety of data sources organizations want to analyze – from applications, social media, and machines, just to name a few – has grown exponentially. The problem is when we bring in all this data, it gets murky: It lacks context and meaning. To effectively analyze the data, we must first clean it, pulling out only the most relevant pieces based on the insights we want to discover.
Once the data is cleaned up just enough to where we can share it, it becomes information. Information is the absolute minimum you should offer a user in a BI dashboard. This means that, if a user is familiar with your business, they would find this data valuable. But in this case, the data only has value if the user knows how to interpret it, and they still need to do work in order to make sense of the information.
An example of this would be a chart showing sales for the current quarter. A sales manager could look at that chart and understand whether or not the sales are good, because they know their goals and what they have done in the past.
The danger, though, is when a user is not familiar with the business. They may look at the sales data and interpret something differently because they don’t have the necessary context – and that’s not a situation any business wants to face.
To add that context, we must take our information to the next level: knowledge. This is where the dashboard designer’s role is essential. Offering users knowledge means that everything has been thought through and they’re not just seeing data hanging in space.
Doing this successfully takes discipline: Never show a number without showing a trend; never show a trend without showing previous trends; and never show those trends without correlating why they are taking place.
For example, don’t just show a single sales trend line. Tell a story with the information by displaying multiple visualizations: last year’s sales, this year’s sales, and any other related data such as goal lines, predicted sales, or a new product launch that resulted in an unexpected spike.
But it also goes beyond simply offering multiple visualizations to show trends. Knowledge differs depending on the user and their role in the organization. The designer’s job is to move across the gaps in the organizational chart in order to deliver knowledge effectively.
One Size Doesn’t Fit All
When it comes to BI dashboards, one size does not fit all. Key performance indicators (KPIs) vary by user, and so do skill sets and roles. It’s important to consider all your users and the type of experience they want to have with the analytics.
Let’s consider an example that an insurance company might need to tackle. They may have two user groups – claims managers and premium managers – that need to analyze demographic information in relation to policies to maximize performance visibility. However, even though the underlying data is the same, the story told by the data should differ depending on what each user is focusing on – either policy premium performance or claim behavior. Complicating matters, a third user – the “line of business” manager – also needs to analyze the data to gauge overall profitability.
When developing the dashboards, here is what I would consider: Premium managers need to answer the question, “Are products and policies priced appropriately?” Therefore the KPIs represented on the dashboard need to include new policies written, as well as the rate of policy renewal and lapse. Managers could then drill down into more detailed information by selecting specific data points, such as a particular agent or product.
On the other hand, the claims managers need to answer the question, “Where are unusually high claims originating?” This dashboard would include KPIs like the number of claims submitted/settled/paid, claim turnaround time, and paid and settled ratios.
Finally, line of business managers who want to understand the overall health of the business, they need to know where the business is succeeding or struggling. Their dashboard would answer the question, “Which channels are performing well and which ones are performing poorly?”
Each of these dashboards is built specifically for the decisions the users need to be making. It wouldn’t do the claims manager any good to look at the premium manager’s dashboard. Moreover, one dashboard that includes all the possible information that anyone in the business could ever need would be overwhelming, leaving users to struggle to find the information they need to make their particular decisions.
Talk First, Build Later
So how can we get here? Before you implement personalized dashboards, it’s important to understand what your users want to get out of the data. You have to remember that you are speaking to an audience – and that audience wants to see something which informs them, moves them to action when necessary, and reassures them when all is well. If you can get a tight focus for your audience, then you can narrow your story and tailor it to meet their needs.
Your users likely know better than you what they want to see; the trick is to get them to put it into words. Remember, you want to represent data in visualizations, not just show them lists and numbers. There are many ways you can show multiple-variable data, from charts and drilldowns to dense information displays. Show users your dashboard and visualization concepts. Ask questions such as: Where is this working for you? Where is it not? Is this solving your problem?
It’s very rare that your first cut becomes the final version, so prepare your users for iterations and collaboration. Bottom line: Talk and draw first, build later. The more knowledge you pull out of your users, the more relevant your visualizations and dashboards will be – and the more easily users will be able to harness data to gain new insights. Transformation complete!
Chris Valas, Senior Director of Program Management and User Experience at Logi Analytics, has over 30 years of engineering and program management experience. Prior to joining Logi Analytics, Chris was the owner of Superient Consulting, where he helped Fortune 100 companies execute technology projects and manage product development. He has also served as VP of Engineering at Brivo Systems and Digital Harbor, and Senior VP of Technology at EKO Systems. Chris holds a Bachelor of Science in Electrical Engineering from Michigan State University and Masters of Science in Electrical Engineering from Northeastern University.
Subscribe to Data Informed for the latest information and news on big data and analytics for the enterprise, plus get instant access to more than 20 eBooks.