This blog post is intended to provide guidance in building great dashboards that will be used in the long term within an organisation. It is based on our experience of building dashboards with our partners over several years, so we now have a good understanding of what works well and what doesn’t.
What are you building dashboards for?
Before building any dashboards, you need to ask what the business targets are that you are measuring against? It is an essential step that ensures that dashboards are adopted and used over a long period of time – not just whilst they are shiny and new for the first few weeks. We have seen many instances where adoption over time has fallen dramatically because the dashboards are only delivering short term value.
For example, it’s easy to create a dashboard for travel expenses based on the department P&L and show travel expenses for year to date and for this period. However, the value comes from knowing whether this is an upward or downward trend, and comparing this to last year’s pattern. For example, are expenses always higher in certain months of the year, because of trading peaks? Or is a particular user way over budget compared to last year without any good reason?
Take the time to understand what the really useful insights are that you can give to your users from the data that is held in your systems.
Context and Change are key considerations
As a first priority for any data dashboard project, you need to understand what questions you need answers to. This will then drive the development of dashboards that can answer these.
An example from HR – a company’s absentee rate is 5% which is within the company norm – but what if all of the absent days for a particular person fall on a Friday – or on a Tuesday after a Bank Holiday Monday?
Make sure that your dashboards do more than just display easy to find data. The context of the data and changes from the norm are key elements to consider.
Is the data in the right format for business reporting?
Often data from source systems doesn’t answer the business questions that you wish to get answers to. This is usually because the source systems are focused on transactions and therefore the data tends to be transactional too.
With our partners’ pre-built datamarts, this problem is avoided. The application reporting datamarts are designed to provide additional business meaning, above and the beyond the transactional data.
For more information on this topic, we would recommend a read of: Eckerson Data Lakes
Intuitive's Top Tips for Creating Really Compelling Dashboards
1. Beware of Averages
Averages can mask the extremes – within the range that is displayed there can be some really high or really low numbers that business users need to be aware of to take corrective action. For example, the overall monthly utilisation of a type of device across an organisation might be 50% (which could be fine), but the figure might mask an underlying under-use or over-use of an individual device by one department or even an individual user.
2. Think about the data
We have explored this above, but this really is the crux of building great dashboards with longevity – what measures, metrics and KPIs do we need to give meaningful business information and measure business performance?
3. Design for Different Roles
The data needs to be relevant for individual users. Make sure that your dashboards provide appropriate detail for the different levels of user within your organisation – for example a team leader might need to see different data from a regional manager
4. Help Users to Navigate the data
Provide a simple pre-programmed route through the data to answer the agreed business questions. Don’t rely on users to filter or use sliders – make it easy for them by pre-building a route that is appropriate to them.
5. Don’t build for mobile
Despite lots of upfront interest from senior management in mobile dashboards, in our experience, very few mobile dashboards are in active business use. And if you have a few mobile users (field account managers for example) then build them a set of their own dashboards designed for the mobile interface.
6. Make it easy
Users don’t want another application to log in to – if they have to they might not bother. So make the dashboards part of your application / web portal or intranet log in.
7. Integrate with standard security
Use the security of your host application to determine the appropriate level of data to provide for your dashboard users.
NB. This is good practice for data governance in this post GDPR environment too.
8. Horses for Courses
If you need to provide row level detail on individual transactions (such as invoices in your accounts system) then provide a drill-out to the host application for the detail. It’s pointless including this level of detail in a dashboard.