How Dashboard Trees Work and Why

...and how to build them

We live in an age of information explosion. Every day more than 300 million terabytes of data is generated. With better editing tools, more and more people have become creators. The same has been happening inside organizations.

But instead of trendy dance videos you see an ocean of dashboards. Thousands of them strewn about every corner of the organization with many created daily.

Despite it all executives still don’t feel like they know what’s happening in their businesses. Instead of getting more insights they get more confusion.

Why is this happening?

In the previous post we covered the problem in detail and introduced Dashboard Trees as the solution.

Let me summarize it here:

• Managers, executives and stakeholders keep asking more questions, which often leads to requests for more reports, more analysis, more dashboards.

• The dashboards themselves feel like a never ending quest. The more features analysts build the more features users ask for.

• With thousands of dashboards, managers aren’t sure which one to use. Often there’s a lot of duplication, the same metric showing up differently in multiple dashboards causing not just confusion but ineffective and incorrect decisions.

• So in order to clarify all this confusion managers ask more questions, demand more reports, analyses and dashboards leading to a vicious cycle.

The consequences

This dashboard sprawl is not free.

There are obvious operational costs associated with it: cloud computing costs and data storage costs. Many of the dashboards require long-running, expensive SQL queries executed daily (or even hourly).

While this was not a problem up until recently due to low interest rates and very low costs of capital, the trend is reversing. Many companies are starting to reduce data teams spend and budgets.

However the more insidious costs are not monetary in nature.

With data teams overloaded with requests and lacking a good way to prioritize the most important ones, many analysts and data scientists feel disillusioned from a career that a mere decade ago that was called “the sexiest job of the 21st century”

Dashboard trees to the rescue

Dashboard trees can fix many of these problems, here’s how.

We know that managers want to make impactful decisions. In order to make impactful decisions they need to see how their domain fits into the big picture otherwise impact will be elusive.

Dashboards, by their very nature, represent only a tiny fraction of the full picture like a small piece of a very large jigsaw puzzle.

Unlike the jigsaw puzzle, managers have no idea what the big picture looks like, so they keep asking for more pieces while trying to assemble them into something coherent.

But the jigsaw puzzle remains elusive.

What is needed is a way to synthesize all this information into a mental model of the business that would allow managers to make sense of the chaos.

Self-serve analytics does not solve the problem. It kicks the can upstream creating more work. Instead of making decisions quickly, managers are asked to do their own analysis and build their own dashboards. That’s like being asked to build your own car when all you want is to go somewhere.

Dashboard trees represent exactly that jigsaw puzzle fully assembled.

Here are just some of the benefits:

• Dashboard trees synthesize the sprawling mess of dashboards into a single coherent big picture that provides much needed visibility into the entire business.

• The hierarchical tree form effortlessly curates information from a high-detail, high resolution level to a 10,000-foot, big-picture resolution level; from the tactical to the strategic in a single view.

• You can finally delete all the other useless dashboards saving hundreds of thousands in cloud costs per month without losing fidelity.

• Once you start using a dashboard tree, and managers see the value, it becomes the single source of truth for the business and the go-to tool for the executive team to run business reviews, conduct planning, forecasting, etc.

• The data team is no longer inundated with requests which leaves them free to explore how to grow the business, which makes for satisfied employees, which means more value.

So how do you build it?

You can start with existing dashboards and reorganize them or you can start fresh by building a metrics tree first. I’ll cover this in more detail in the next issue but until then here are some examples.

If you want to build a metrics tree first, start from the top metric you care about. Here’s an example of top line metrics that would typically fit into the L0 level of the tree above. This is from an actual client. You’ll notice that Live Bookings is their goal.

Then you continue building metrics trees using all the available metrics in the business. At this point they will be disconnected, but that’s ok.

Here’s an L1 metrics tree for New Business Booking which can impact New ARR thus connecting L1 with L0.

Here’s another metrics tree for upsell that could fit in the same L1 dashboard.

Next you connect all the metric trees and separate them into zones that form the levels.

If you want to start with organizing existing dashboards, you can start with the customer journey and roughly follow the org structure. It’s important that dashboard trees not follow org structure too rigidly because it changes.

As we’ve worked with more clients (at Wave40, the agency I consult for) we’ve noticed for example that you cannot neglect the power of that structure if you want to succeed in implementing dashboard trees.

There’s a lot more to it of course, and you need help building this for your org feel free to reach out to me directly. In the meantime I’ll also take any questions you might have.

Until next time