The dashboard is where most end users will consume the data. Assuming you already have accurate, timely and comprehensive data, the job of designing a dashboard comes down to selecting the right metrics and presenting them in a clear manner.
Before you start, ask yourself the following:
- Who will be using this dashboard?
- Why do they need the dashboard?
- How are they currently meeting this need?
A modern data organisation typically aims to make data widely available, and you will need to keep this in mind when preparing the dashboard. The primary user might be a single person, but keep the wider organisation in mind too.
The team in question might have a set of OKRs they are measuring their progress against, it can be used to highlight issues in an operational process, or to measure SLAs. If your organisation is going through a data transformation project, it’s also possible that the team isn’t sure what they need the dashboard for.
The current solution might be a combination of spreadsheets and tools. Replacing existing solutions with a dashboard will require careful listening and observation, and you might not get to the bottom of it without first presenting a solution to the team.
Once you have answered these three questions, you are ready to start designing the dashboard. If you’ve done your job well in setting up the company’s data infrastructure, you already have timely, accurate and comprehensive data available to make use of.
Your main concern will be to define the right metrics, and present them in a compelling way.
When selecting metrics, here are some questions to ask yourself:
- What is the nature of the thing being measured? Is it a step-by-step process, such as sales or user onboarding. Or is it an ongoing process, such as payment processing or SLAs? If it’s financial data, you might want to present data in terms of months, and year-to-date. If it’s product data, daily or hourly data might be relevant.
- What does successful use of the dashboard look like? What actions will the user take that will be reflected on the dashboard?
- How do you ensure that the metrics are comparable across time and other dimensions?
When preparing the layout, consider the following:
- Based the layout on the structure of what’s being measured. If it’s a step-by-step funnel, the first chart should be either the start or end of the funnel.
- Keep things as simple as you can. A stacked bar chart that shows the metric over time is usually the most valuable. Line charts are useful when a chart contains several units, such as different currencies.
- Show one metric per row. Include several charts of the same metric where appropriate, such as daily, weekly and monthly views.
- Add any especially important single-value metrics at the top, such as last months revenue.
- Add titles and descriptions to the chart where appropriate, but use it sparingly as keeping these up to date might be a time consuming task. Dashboarding tools such as Metabase make it easy to add metadata to tables and reports that can be easier to maintain.
Once you have created your dashboard, the real work begins. You need to make sure that the primary user understands the dashboard and is ready to take ownership over the numbers there. Workshops and training sessions are useful to get everyone involved. You will also need to make sure it continues to show accurate numbers. Many dashboarding tools offer analytics into dashboard usage, which tells you if you’ve been successful in introducing a new dashboard to the organisation.
Leave a Reply