Decisions using dashboards - Part One

We have spoken to 25 business leaders so far, in terms of how they use dashboards and make decisions using data. Here are some key insights.

We are about a third of the way in terms of understanding how people and businesses make decisions using data in general, and dashboards in particular. We’ve done about 25 conversations so far, and that provides sufficient “data” for an interim summary on how people use dashboards.

So here goes:

1. Senior leadership seldom uses dashboards: Now the reasons for this are many, and vary across all the people that I spoke to.

   1. On the supply side, most standard dashboarding tools have suboptimal mobile experiences, and most senior people operate exclusively on mobiles

   2. On the demand side, a dashboard is yet another website / app that the executive needs to go to. Far easier for them if they can get the data through their normal means of consumption, such as email or slack. In most companies I spoke to, it is a common practice to copy paste data insights and graphs into email / slack.
      
      Many companies have configured alerts such that a snapshot of the dashboard gets emailed to key stakeholders periodically (sometimes as frequently as hourly).

   3. Sharing data manually allows people to massage the numbers, and they like the control.

2. Top-down mandate helps in terms of being data-driven: In the few companies that I spoke to where everyone actually uses dashboards (or data) for decision-making, this is usually a function of a top-down mandate, where the CXO insists that everyone use the dashboards.
   
   A corollary of this is that sometimes tools get to be associated with certain individuals, and when there is a leadership change, the old tool gets discarded.

3. Many companies use multiple dashboarding systems: Sometimes this is a function of certain tools becoming specific to certain use cases (such as HubSpot for marketing). In other cases, decisions are left to individual teams who use the tools most convenient to them.
   
   One downside of this, of course, is that the data in different dashboards may not agree with one another (I’ll write separately on this data consistency thing - it needs sufficient attention), and reconciling this can be a time-consuming (and willpower-consuming) task.

4. Comfort with data is a function of when someone started their career: People who started work in the last 15-20 years are far more comfortable with data (on average - obviously I’m stereotyping here) compared to those who started their careers in a more data-poor environment.
   
   Another way of saying this is it’s not just companies that can be “digital native”. Even people can be

5. Dashboards can be heavy, and take minutes to load: This, according to some interviewees, was a dealbreaker in terms of using them. “You open the dashboard; by the time it shows you the data, you think you can check some email, and you never return to it”.

6. Organisations have a lot of “dead” dashboards: Metrics go in and out of fashion. Business changes. What was once a novel pattern may not be novel any more. Markets and customer behaviour might change. All these mean that dashboards have a short shelf-life, but are seldom decommissioned.
   
   In my own experience, I remember trying to shut down some “dashboards” (more like email alerts) in my previous job, and every time I tried to do that, there would be some manager somewhere who would write back saying this dashboard was critical to their work!

7. Senior leadership doesn’t really want interactivity: It is rather common for analysts to provide lots of interactivity in a dashboard, with the intention that it can cover for more edge and corner cases, and give more power to the leadership to see the data.
   
   However, leaders seldom find the interactivity useful, and instead only want to be shown what is useful / interesting (and hence all the copy pasting into email / slack)

8. Dashboard tools is a platform business: Market leaders such as Tableau and PowerBI have a headstart in that they seem to have cracked the two-sided market of customers and developers. So people prefer one of these because it is relatively easy to find people to build dashboards, and analysts learn these tools because more companies use them.
   
   So, while one may not intuitively think of it that way, dashboarding tools are also a “platform economy”, and that explain why we have some dominant companies in an otherwise fragmented market

9. Companies use dashboards as ETL tools: This was surprising when someone first mentioned it to me, but then more people said the same thing. Dashboards are used as single sources of truth. In more than one company I spoke to, the only source of data analysts can use for their bespoke analysis is PowerBI.
   
   Another interviewee said, “50% of the use case of a dashboard is to download data back and analyse in Excel”!
   

10. PowerBI doesn’t offer good built-in visualisations: This is specific to the tool, of course. They have a marketplace where you can buy add-ins for specific tools. However, in large companies with large procurement departments, this can take a very long time.
    
    In my opinion, Tableau defaults aren’t great as well. That said, in a corporate context, I’m not sure of the return on investment on “good visualisation”.

11. RoI in looking at dashboards can be low: Metrics hardly change, or change very slowly. From this perspective, there is often very little information in looking at dashboards. This is another reason senior people don’t use dashboards.

12. Customisation is a problem, both ways: Having a centralised BI team means when a particular team wants to track a custom metric, it is difficult to get it added to the dashboard.
    
    On the other hand, when it is easy to get a team’s custom metric added to the dashboard, it (the dashboard) can become rather heavy and noisy, making it harder for people to glean insight from them.

13. Dashboards are not optimised for mobile: This is worth repeating, given that senior management seldom uses computers. Standard dashboarding tools were built for a computer-first experience, and mobile has been added as an afterthought.
    
    This has a few implications - dashboards are heavier and denser than they need to be; fat fingers means interactivity on mobiles is hard (so in the worst case you need to give good defaults); design that is optimised for computers may not look that good on mobile; etc. etc.

How do you or your company use dashboards? Is what I’ve put here representative? If not, I’d love to know your views. You can either leave a comment here, or simply reply to this email, or let me know if you are willing to have a more in depth conversation (~30 minutes) on this topic.