by W. Scott Roberts

What are software dashboards?

Software dashboards are visual displays of metrics that provide real-time information at a glance. They summarize the performance of areas within a company, department, or system.

Dashboards go by many names: executive dashboards, executive reports, scorecards or control panels.

What makes a good software dashboard?

The popularity of business intelligence software has made dashboards easy to create and quite flexible. However, it remains fairly difficult to create effective and useful dashboards. Too often they are cluttered, flashy and downright uninformative.

The mark of an effective dashboard is its ability to convey information quickly and clearly without requiring the user to spend a lot of time deciphering the contents.

That’s why using the metaphor dashboard is so effective: most people already know what a dashboard is used for in an airplane or an automobile, and they know how to read a dial or gauge to learn how much fuel is available, the temperature of the engine or how fast they’re going. Using the same indicators in a software dashboard means that most people already know what they’re looking at with regard to the representations of values – they just need to know what the representations are telling them.

Designing a dashboard

It’s important to plan out your dashboard before you begin building it. Start by answering the following questions: Who will use it? What is their role in the organization? Will they need one dashboard, two or several? What is the best way to present and arrange the information they need to see?

Dashboards usually offer a variety of choices of visual representations of metrics. You should decide which type of graphic element will best represent the data the intended audience needs to see.

1) If you have a single metric you wish to display, consider using a text box to display a numeric value or a gauge to represent the value graphically
2) Charts and Key Performance Indicators (KPIs) are best for showing trends in a limited series of data. Charts are good for comparing two series; KPIs are good for comparing values to a target value
3) Use a grid in conjunction with a filter (to limit data) or a Daily Monitor (to change datasets) for large amounts of data.

What to avoid when creating a dashboard

Try not to waste space in a dashboard. The point of the tool is to present a lot of information in a small area, make efficient use of it.

Don’t represent data with inappropriate devices. A dataset with eighty elements isn’t going to be easy to read as a pie chart but might work great in a grid. A small set with five elements and a target value won’t be best represented in a grid, but is perfect for a KPI.

Since the human eye is naturally drawn to larger and more colorful visual elements on a page, be sure not to emphasize the wrong things by making graphs or charts too large or flashy. A good practice when creating dashboards is to decide at the start which data will need to be featured strongest. That way you can be sure to draw the attention to the most important information right off the bat without the need for a drastic redesign.

Dashboards in PresentationCenter

Precision.BI’s PresentationCenter performs best when you limit the size of Worksheet and Analysis (crosstab) objects to less than five megabytes (5mb). This economy of design will allow you to add more source objects without bloating the total size of the Presentation Access Center (PAC) (note that you can check the size of your sources and PACs in PresentationCenter Designer).

Objects of five to thirty-five megabytes will also work but you should take care not to add too many to a single PAC, lest the total size become too large and the PAC too slow.

Adding objects of over thirty-five megabytes should be avoided; those over fifty megabytes may cause “Out of memory” errors and are not recommended.

Adding custom report objects to the Object Panel has it's own set of rules; less than ten megabytes is the optimal size, ten to twenty megabytes will work but may be slow, and those over twenty megabytes are not recommended.

Again, remember that with each object you add to a PAC, the larger the overall size of the PAC will become. Extremely large PACs will be very frustrating to use, since they’ll be slow and may even generate errors – and those issues would certainly be in conflict with the purpose of a dashboard in the first place: to see accurate information displayed quickly!