Death By Dashboard…and How to Avoid It!
Drag and drop, slice and dice, sort and filter, drill-down…these are just a few of the many terms people use when talking about dashboards. A dashboard is a tool that displays metrics in a graphical and tabular format, typically via an interactive system that allows the user to view important information on single screen or page. Dashboards often provide the current status of key performance indictors (KPIs), and the comparison to historical trends or pre-set benchmarks so the user can quickly identify whether their organization or team is on track to achieve objectives. In theory, dashboards sound like an effective and efficient way to document goals, track progress, and identify misses before they occur. However, many of the reasons people are attracted to dashboards are the same reasons they can be ineffective at driving action. I summarize these double-edged swords below, and provide recommendations for organizations that may have already built a dashboard or are thinking of doing so.
Many of the reasons people are attracted to dashboards are the same reasons they can be ineffective at driving action.
They provide a lot of information. Often too much! Dashboards often contain so much information that they lead to information overload and analysis paralysis rather than true understanding and action.
They allow for slicing and dicing. End users often spend so much time “playing” in the data to satisfy their curiosity that they fail to answer any relevant business questions. The interactivity of a dashboard (slice, dice, sort, filter, etc.) can also lead to such small sample sizes that the insights can no longer be trusted to generalize back to a larger population.
They allow for easy comparisons. Dashboards often provide comparisons over time and across groups; however, they typically do not include significance levels or effect sizes that tell you whether the differences are significant, or large enough to matter. This often turns into incorrect assumptions or interpretations of the data, and perhaps, unnecessary action.
They allow for self-service. Many organizations implement dashboards to reduce or eliminate the need for static and manual report generation. However, self-service can turn into “no service” if business leaders do not have the capabilities, interest, or time to use the dashboard.
They allow for exploration. Although data exploration may lead one to identify unexpected problems or issues, it can also lead one to “find” issues that aren’t really there. End users who use dashboards to “dig around in the data” rather than test hypotheses or assess KPI status run the risk of finding something they believe to be significant when it is actually just due to chance.
Already built a dashboard? To ensure that the right people view the right information in the right format and take the right action, I recommend building your dashboard around a limited number of important KPIs that have set targets or benchmarks. This will ensure your end users understand what they are looking at and why. I also recommend that all users be trained not only on how to access and navigate the dashboard, but also on why the metrics are important, how to interpret findings, and when to take action based on the results.
Haven’t created a dashboard yet? I recommend that you take time to determine the main questions you are trying to answer with your data, and the end users who will be responsible for taking action. Once you understand your questions and audience, you can determine whether a dashboard or another delivery mechanism (report, scorecard, data story, etc.) would work best.
I am interested in hearing your thoughts on this topic! Feel free to post your comments – whether in agreement or opposition – in the comment section below. Also, feel free to reach out!
