Rethinking Dashboards for HR Analytics

dashboard

Just because you provide users with “analytic evidence” doesn’t mean they’ll believe it, and in some cases, it only serves to cements their current view or bias. Here is a key lesson: Data analytics and visualisation is a creative activity, and unfortunately not everyone is going to like or appreciate your artwork, writes Rob Scott. 

You may be surprised to read, given my appreciation for technology, that I’m not generally a fan of dashboards for analytics. I don’t mean the actual software platforms, there are many great products available, but the way these tools are often used is outdated and more aligned with times when hierarchy, standardisation & control were the dominant drivers of business success.

Modern analytics is about creating a process which translates patterns of information into actionable insights, and with enough context.

Single-screen dashboards were an easy way to represent predictable environments using easily defined and limited KPI’s. But the modern digital organisation operates quite differently. Change and agility are constants, and people have tremendous freedom to operate creatively to maximise their personal contributions.

We also generate and have access to lots of additional data. Together with AI technologies that help optimise our decision making, the idea that a single-screen with complex and highly interpretative visualisations is a panacea for achieving our business goals, seems a step too far.

Many of these new dashboards look impressive, designers are being applauded for feats in artistic design and coding, but your users are likely to struggle to find insights they need to deal with their business issues, opportunities and innovations.

Why is this the case?

Modern analytics is about creating a process which translates patterns of information into actionable insights, and with enough context. These insights shouldn’t just be “interesting & cool stuff” – that only serves to distract users.  Rather, analytics must firstly relate to a business problem you are trying to solve and secondly should produce personal “aha moment”. In other words, insights should create a realisation that a current position, view or belief you hold is wrong or no longer valid.

Just because you provide users with “analytic evidence” doesn’t mean they’ll believe it, and in some cases, it only serves to cements their current view or bias. Here is a key lesson: Data analytics and visualisation is a creative activity, and unfortunately not everyone is going to like or appreciate your artwork.

The fixation with single-screen dashboards, crammed to the rim with important, but often unexplained content, coupled with an allure to miraculously find hidden secrets in data, has inadvertently pushed developers and analysts to aggregate results in order to save space, resulting in interesting “averages” but little insight.

To compensate for this, dashboards platforms offer features like filters, selectors and no-code scripting, which requires a user to interact with the data and search beyond the averages for insights.

Many of these new dashboards look impressive, designers are being applauded for feats in artistic design and coding, but your users are likely to struggle to find insights they need to deal with their business issues, opportunities and innovations.

From an end-user perspective it becomes too complicated. Using the dashboard requires time and practice, and users may also find it difficult to replicate outputs because the underlying data is dynamic.  So, they simply stop using your creation and much of your effort is wasted.

My top 5 points of advice are:

  • Be clear on what you are trying to achieve with analytics
  • Displaying metrics and building analytic insights are two different things
  • If you want to digitise your monthly HR report, show financial trends, or call-centre metrics, single-screen dashboards are a suitable medium
  • If you are trying to solve business problems using data, a pre-designed dashboard of is not the best solution. It supposes you made significant assumptions on unknown outcomes
  • Rather make each business problem a bespoke analytic project, use skilled data scientists to collect, process, structure and visualise the right data, and then work together with your analyst to translate the results into meaningful business insights.

Image source: iStock