Back to overview

Dashboards, Big Data, and KPI’s: A Roundtable Discussion

Dashboards that bring data to marketers’ and researchers’ desktops are all the rage. But without clear planning and execution, dashboards can become sources of confusion and distraction.

Three GfK experts who design and work with dashboards from different perspectives recently sat down to share ideas and insights: Bitsy Bentley (Director, Data Visualization), Martin Ho (Vice President, User Experience), and Florian Kahlert (Manager, Digital Market Intelligence). Here they discuss dashboard users, problems with off-the-shelf packages, and whether Big Data is driving the dashboard revolution.

What is the best way to approach dashboard planning and design?

Ho: You unpack the problems clients are trying to solve.  Often the goal is simply enabling users to access and consume information really quickly. In many cases, we actually want users to spend just a minute or two glancing at the dashboard.

Kahlert: One of the challenges clients face is satisfying multiple constituents within an organization. You have the researcher, who needs a lot more in-depth information; and then you have the end business user, who may or may not have a need for detailed tables and just wants to see charts. One of our clients, the president of a large advertising network, said to me, “I just need charts that look nice – red, blue, and yellow – that I can copy and paste and send to my clients.” So we try to talk to all constituents before planning anything.

Bentley: We find that there are so many varieties of visual and statistical literacy that we need to work with. The primary purchaser of these systems and programs tends to be a very senior analytics person who is super concerned about data validity and wants to know how big the base sizes are. But the broadest group of users – the colleagues of the analyst – often has no idea about statistical conventions. And so they tend to focus on things like color.

What about off-the-shelf dashboards? What is out there, and how well do they work?

Bentley: The tools are only as good as the people who are using them. If the thought process behind how these dashboards are put together and applied is incomplete, the tool is not going to be very useful.

If you want to have a really good dashboard experience, you’ve got to have a really good front-end developer. That can be challenging in market research, because we tend to focus so much on back-end data management. But we’re starting to see a shift away from that emphasis, which is really exciting.

Ho: We have heard from some clients who have used off-the-shelf dashboards that they had to pigeon hole the way they were thinking into the functionality and analytics offered by that package. In reality, they just wanted an optimal way to get the data they needed to summarize or analyze.

 To what extent do you think the need to access and analyze Big Data is driving demand for dashboards?

Bentley: The biggest barrier to unleashing the potential of data is understanding why the information is relevant to you. What is the business case for actually doing this examination? That, I think, is the piece of the conversation that sometimes gets overlooked, because people become so focused on the huge data resource that they have. Leveraging Big Data is truly effective only when we have identified the problem that you are trying to solve.

Kahlert: The amount of data that clients have is driving the need for more analysts who can pull the data out and make sense of it. The questions that are relevant for business decisions aren’t going to change. You always end up managing the main elements of your business on a few key performance metrics. That data may be buried in the client’s servers, and you need an analytics guy to find it.

Ho: Sometimes companies think that dashboards are really going to be the silver bullet, and we have to re-orient them to the idea that dashboards are just a starting point. The most productive or professional solutions are ones that enable them to take the shortest, most practical path to where they need to go.

Bentley: What’s really driving the desire and the demand for dashboards is that we finally have the computer processing power to do these kinds of calculations on the fly. And this is where effective data management comes in.  The number one barrier to good data visualization is good data management.  It’s like 80% data management, and then 20% the layer that you put on the top.

What do you see as the future of dashboards? Where are they going, and where would you like to see them go?

Kahlert: Tying the dashboard to external actions is going to be critical.  If the dashboard tells you something is wrong, you usually have to go out of the application – or even to a different computer – to investigate further or start fixing things. Being able to integrate the top, relevant tasks coming out of a user’s dashboard directly into that interface – I think that’s where things are going to go.

Ho: It really is more than just displaying some things in an aesthetically pleasing way. It is actually shaping experiences, and therefore behaviors. One direction for the future of dashboards is the integration of insights alongside data visualizations. Insights go beyond telling you something is wrong; they can tell you what to do. This removes the ambiguity of what users should do next – which can be particularly useful when dashboards are used infrequently or by novices.

Bentley: We’re actually starting to integrate third-party APIs into the presentation of the data, because that is what clients desperately need. With one client, we’re using three different data sources for a single interface. And that’s a pretty big departure, from a market research standpoint.

Ultimately, I think it’s aboutusing data to create smarter tools for people. In some of our applications, the users’ choices feed back into the database, and we can actually start to do analysis on that behavior and improve their experience. The really core philosophy is: Providing more opportunities for data adoption within organizations. Let’s make it easier for them; let’s make it smarter.

Back to overview

Write a comment

*required field

Leave a Reply

Your email address will not be published. Required fields are marked *

Your comment*