This isn’t a case of dashboards being good or bad. The reasons that dashboards fail at these newer use cases is because they’re so far beyond what dashboards were designed for
. It’s like buying the best plane on the market and then expecting it to fly you to the moon. Whilst superficially similar, the use cases are so dramatically different that no amount of modification will make it fit for purpose. You have to start with a completely different set of technology.
Where do the flaws come from?
Dashboards were originally designed to monitor a business and they’re a great tool for doing that. They provide an overview into the how a set of metrics are tracking and what some of the trends are. But they’re a terrible tool for data discovery.
This issue has been compounded by the increasing amount of data that sits within every business. When people first started building dashboards, there wasn’t a lot of data so it was easy to build a dashboard that showed the key metrics. Now, there’s so much data available we can’t fit it all on one page but we keep trying to - because that’s the promise of the dashboard.
To make everything fit on a dashboard we end up aggregating the data, so by the time a business user sees the dashboard the data has been averaged away. In many cases, the top line number gives no visibility into what's happening in their underlying business.
Nate Silver said it best in The Signal and the Noise: Why So Many Predictions Fail - But Some Don’t
, “Partisans who expect every idea to fit on a bumper sticker will proceed through the various stages of grief before accepting that they have oversimplified reality.”
To solve this problem vendors have added filters and drill-downs which push the onus of discovery back onto the user. Users have to slice and dice the data themselves to find insights. This gives rise to three problems.
1. It’s not repeatable
Data discovery becomes entirely accidental. If a business user finds an insight today, the chances of them being able to go through the steps to discover the same thing tomorrow are minimal.
2. Analysis fatigue
Even if a business user does manage to repeat the process of data discovery, they may have to keep going back to check because the data doesn’t change every day. If you’re continuously looking at something that fundamentally doesn’t change you eventually stop looking at it. This analysis fatigue means that users may miss finding out when something does happen.
3. Human bias
If a user thinks they know the business, they may not look at some things because of human bias. People don’t have the time or capability to dig deeper into everything, so they have to choose where they look. For example, if a manager trusts a particular salesperson, they may never drill down into that person’s business to see if they’re going in the right direction.
If a user wants to find new insights about their business then they need a solution that can search through large volumes of data and identify trends. That’s a very different use case to what dashboards do well (monitoring the changes in a set of metrics).
Vendors don’t want to talk about this
The problem is that most vendors don’t want to talk about the flaws of dashboards. From one perspective, it’s understandable – Tableau, Qlik, PowerBI (and yes, Yellowfin) have put more than a decade into building the best dashboard products possible. But from the perspective of the user, there are a lot of use cases that would never be delivered with a dashboard from first principles. As Abraham Maslow famously said, “If all you have is a hammer, everything looks like a nail.”
All of the major BI vendors have a vested interest in dashboards. It’s what they sell. So they continue to push every possible analytical user experience into that paradigm rather than looking at the flaws of dashboards.
At Yellowfin, we’ve recognised the flaws of the dashboard over many years. They simply don’t solve every problem for users. That’s why we’ve been building a solution that automates the challenge of discovering insights. Dashboards will still play an important role in monitoring and understanding whether a business is on track, but automation is much better at understanding the significance of what’s happening underneath.
We’re calling this product Signals, and it will go live as part of Yellowfin 8. We’re incredibly excited about this product and think it’s it’s going to change the paradigm of the dashboard and how business users think about data discovery.