Dashboards are Dead, Long Live Dashboards

We must change how companies use dashboards.  Too many businesses rely on dashboards as the first line of analysis.  For an organization to get the most out of their data efforts, dashboards must take a step back and become a second level point of analysis.  Two major trends have emerged for why companies should change how they use dashboards.

Proliferation

This development comes in two ways, both of which should strengthen your use of data but can also undermine it..  One, you have access to more data than ever before.  Two, it has never been easier to create visualizations of that data and build dashboards.  So where is the fly in the soup?  The problem is you can’t give proper attention to all the data coming to you, even after putting it into a set of dashboards..  Honestly, when did you last spend the time for an attentive review of all the dashboards and reports available to you and pertinent to your work?  Many of us drown in data and reports.  As a result, we get a haphazard sampling of the analysis based on our available time, the somewhat random topic du jour that focuses our attention to a subset of reports, or a crisis that triggers a meaningful review of a single area.  Otherwise, much of the analysis sits on a digital shelf, available but not utilized.

Context

Dashboards do a poor job of conveying the context you need to truly understand the significance of what it shows.  This can lead to wrong conclusions and reactive actions based on misinterpretation.  Most dashboards show a snapshot of data for the week/month/year to date.  Dashboards often include an up/down indicator based on the comparison to the previous week/month/year to date.  The limitations here come from providing a limited sample of data and no broader context.  You have the numbers for that moment in time, but little to tell you the broader trends for the data.  For example, you may be down month over month, but how did last month perform?  If comparing against the best month in the past 5 years, then being slightly down may report “down 7% month over month” when it should read “second highest month in the past 5 years.”

So Where Should Dashboard Fit?

Dashboards still have an important role, but not at the frontline or first step of your analysis.  Your first line of analysis for your business should involve a monitoring or curation solution.  This approach watches all of your significant data points and flags variations and trends worth your attention.  This gives you comprehensive awareness of what drives your business.  When one of these metrics deviates from the status quo the system alerts you that something of interest has happened.  For example, the system may monitor the number of new customers who made a purchase over $200 each week.  If that number fluctuates in a significant way, up or down, you get an alert.  As part of that alert, you get a link to a dashboard focused on a broad set of metrics for new customers.  You also get links to the two or three most likely causes for that shift, such as changes in online activity, overall buying trends across customer types, and other factors.  Now you have the relevant context and plus deeper analysis to accelerate your investigation of the change.  In this role, dashboards provide a broad skim of the surface to further orient you on “where you are” before diving into the root cause analysis.  We’ll explore the difference between a “monitor” and a “curation” system in future posts.

For now, just realize that if dashboards serve as your first line of analysis, you are constrained by your team's capacity and discipline to review the data, they only get part of the picture, and it may be a distorted picture of your business.  To counter these weaknesses, build detection of significant irregularities as your first line of analysis and move the dashboards into a supporting role.  Future posts will explore the approach Winealytics uses to provide this approach to our clients.