Skip to content
Home ยป How To Do Dashboards Right

How To Do Dashboards Right

“We have x data and we would like to build a dashboard with it”

-Companies that consider themselves data-driven

While I wouldn’t consider my friends wanting a dashboard of “all the times you’ve been wrong over past 10 years” a dignified use of my skillset, everyone seems to want a dashboard these days.

At Xabit, we get a lot of calls from potential clients wanting to use our services to build dashboards for themselves and/or for their clients. It’s no surprise that the first thing organisations consider when they think about becoming data-driven is a dashboard. After all, dashboards have numbers (i.e “we have data”) and are something tangible to show to management/investors/clients (hence “we are data-driven”) etc. With no code tools like Tableau, PowerBI and Metabase, it’s also easy to get up and running in a matter of hours, let alone days or weeks.

After we furnish a proposal however, we usually get one of two reactions: “that’s too much for just a dashboard” or “all you’re doing is putting charts on {put tool name here}, why does it cost so much?”. For us, this is becoming an FAQ so we thought it would be best to write a blog post on how to do dashboards right. Here goes.

The first thing we’d like for you to understand is that there are three basic buildings blocks of any dashboard. They are:

a) Key Performance Indicators (KPIs): We like to think of KPIs as metrics that distill aspects of your business down to a single number. For example, Revenue is a KPI that distills your sales strategy down to a single metric, while Average Recurring Revenue is a number that might distill your customer retention strategy down to a single number.

While KPIs primarily help you gauge the health of the most critical parts of your business, what’s ignored is the fact that they also help rally teams around a single goal: improving the KPIs.

b) Dimensions: The second building block of a dashboard is the dimensions (angles) through which you want to follow KPIs. Dimensions are important because they help you break down KPIs into smaller groups that can be compared to each other. Take for example the dimension of time: by breaking down the KPI Revenue into daily dimensions, you can compare yesterdays Revenue against the Revenue today. Similarly, by breaking down Average Recurring Revenue by customer, you can compare your top customers against the bottom ones.

You might be wondering how one goes about finding out which dimensions are important. KPIs are usually broken down either by a measure of time (…day/week/month…) or category (customer groups, product categories etc.). Which exact dimension to use is usually a factor of the business model. For a cafe, breaking the Revenue down by day might be important whereas for a construction company, breaking Revenue down per quarter is perhaps a better approach.

c) Visualisations: Once you know the KPIs and dimensions, you simply need to select the right visualisation to present them. The core criteria when deciding on a visualisation are twofold:

  1. Is the visualisation coherent with the KPIs and dimension I have chosen: for example, a line chart represents revenue growth over time much better than a pie chart.
  1. Is the visualisation easily understood by all the stakeholders: for example, having a animated Sankey diagram looks great, but if you have to write a 1 page document explaining how the visualisation works and what it’s supposed to show, you’ll never get stakeholders to use the dashboard as intended.

That’s it. It’s that simple.

So why would anyone pay any money to follow three simple steps? Well here’s our (FAR) Frequenty-Asked-Response to this question: Is it easy or hard for you to distill your business strategy into a few metrics? Is it easy or hard for you to translate your business model into dimensions? Is it easy or hard for you to figure out which visualisations show metrics and dimensions in a way that anyone looking at the dashboard can intuitively understand your business strategy and model and consequently suggest ways to improve them?

If your answer to the above questions was easy, we agree…you shouldn’t pay anyone any money. You might pay someone like us to show you how to get it right, but you can execute things yourself. If your answer to any of the questions was “hard”, well, you know who to call. P.S – It’s not the Ghostbusters!

On a parting note, notice how we haven’t talked about data at all as a building block to doing dashboards right? In our opinion, starting from the data is the wrong way to make good dashboards. We prefer to work backwards: define the metrics (KPIs) you want to track first, and only then consider whether there is data to build that KPI into a dashboard. Doing so helps us uncover critical gaps in your data. Is it the scientific method? Perhaps but it’s definitely two birds with one stone.

Keep Data. Decisions. Repeat-ing,