Blogging all things data

Create Dashboards people will actually use

Making valuable decisions is one of the most important functions in business. Dashboards are a tool that are used to help make the right decisions. So why are so many dashboards rarely used?

Many people suggest that having a good dashboard design ensures people will use the dashboard. We've found that business people do not use dashboards for a number of reasons:

  1. They don’t know the dashboard exists or how to access it
  2. They don’t know how to use the dashboard to extract the information they require
  3. The information displayed on the dashboard is not the information they need to support their decision making
  4. The information in the dashboard is out of date
  5. The information is presented in an undesirable manner
  6. The data is not correct or it doesn’t match another report 

How can we ensure that we are creating a dashboard that will be used? 

Make the Dashboard easily discoverable and accessible

The first thing we can do is make it easy for a business user to find and access a dashboard. One way we can do this is to provide a centralised portal from which all dashboards can be accessed. This eliminates the need for a business user to navigate numerous applications to determine whether a particular dashboard exists. If all dashboards are registered in a centralised portal, then the dashboards are more easily discoverable by business users. 

Provide training on how to use the dashboard

Not all business people have been exposed to analytical reporting tools that a dashboard may be authored with. As such, it is imperative that people are provided training on what functions are available, and how to use them. Dashboard Consumers should be taught how to perform functions such as filtering data, drilling down, saving views of data, exporting data, and slicing and dicing interactive dashboards. If a Dashboard consumer does not know that these functions are available, or how to use them, the dashboard may be deemed useless to them. For example, imagine a dashboard that tracks the entire company sales. If a salesperson responsible for a particular region, does not know how to filter the information on the dashboard to only show sales for their region, how useful would the dashboard be to them?  

Ensure the dashboard displays the right information in the right format

Know the audience and purpose of the dashboard you are designing. Avoid making invalid assumptions about what you think your audience may need to know.  Workshop the requirements of the dashboard with the intended audience. Keep in mind, you have a limited amount of space to convey information. It is critical that the information you chose to display is relevant. Focus on the information and metrics that matter.

Then design the dashboard in a format suitable for the audience.

There are 3 types of dashboards:

  1. Strategic / Executive Dashboards (Scorecards)

These types of Dashboards are usually targeted towards the Executive level audience. They usually contain KPIs that are relevant to measuring the success of a business process or objective.

  1. Analytical Dashboards

Analytical Dashboards are usually targeted towards Analysts and Middle level Management. They typically contain information relating to the drivers of KPIs. Analytical dashboards enable users to slice and dice drivers to drive insight.

  1. Operational Dashboards

Operational Dashboards are usually targeted towards front office staff. These dashboards don’t usually contain the slicing and dicing capabilities of Analytical Dashboards, instead they are designed in a format that directs action.

Ensure the data in the dashboard is appropriately refreshed

How frequently should the data in the dashboard be refreshed? Take a car dashboard for example, how helpful would it be to show the speed you drove yesterday, rather than the speed you are driving right now?

Also, it is important for someone using a dashboard to have an understanding of how current the data is. For example, a data warehouse may only be updated nightly, but this may not be immediately obvious to a dashboard user. Including a “data currency” field on the dashboard can overcome this blind spot. This can sometimes be trickier than it seems with different data sources available at different times, possibly a topic for another blog…..

Make sure the dashboard looks good

First impressions are important. We only have one chance to make a good first impression. The dashboard must be visually appealing. When designing your dashboard, consider the device from which it may be consumed and design accordingly (e.g. Mobile phone).

  • Use clear, intuitive data visualisations.
  • Don’t make every visual the same size on a dashboard. You need visual cues to highlight what matters the most on a visual. Use colour and size to do this.
  • Create separate dashboards for users that want to know more.

Empower the user to finish their task quickly and understand the action that should be taken based on the results. 

Make sure the numbers are right and defensible

Ultimately, the reliability of the data underlying the dashboard will be the most critical success factor.

Having a test regime when building a data foundation is vitally important.
This includes getting the right subject matter experts to define the business rules and quality acceptance criteria.
Often, the numbers in a dashboard may be compared to another pre-existing report or system screen.

We need to firstly determine that the existing report is correct according to the business rules that have been agreed across the organisation.

We need to then determine whether the report being compared to is fit for purpose for the decision being supported.

If we’ve managed to ascertain a correct reference point with an existing report or a different set of “expected results”, we can then implement a quality control process.

Typically this consists of an automated reconciliation that reports on the accuracy of data being summarised by isolating inconsistencies, missing data and mismatches.

Making these reconciliations transparent helps to establish trust that the data is being regularly checked before being published.

 

Written By:

Principal Solution Advisor: Karen White.