Is a gauge or dial effective for visualizing management information?


A management information dashboard is an interactive device to enable the monitoring and comparison of performance.  To be effective, dashboards should be designed to take best advantage of human visual perception and cognition.

In this business context, the term ‘dashboard’ is an imperfect metaphor. The dashboard of a car is designed to show the driver the status of what is happening right now, rather than trends over time. An analogue instrument such as a speedometer or rev counter will show at a glance whether a crucial measure is increasing or decreasing, or whether it has reached a critical level. It provides immediate feedback to the driver of the effect of changes made to a control such as the accelerator or brake on an important measure such as speed. When driving a car, there is a direct relationship between changes made by the driver through the controls and the measurement that is displayed on the dial. However, in business there are usually many controllers (“business drivers”) that affect the key output measures. There is not usually a simple way to see the effect of adjusting any one business driver by monitoring a single output meter, because of inter-relationships between the business drivers involved and the time lags.

Dials and gauges are generally ineffective for visualizing management information.  There are several reasons for this.

1. A dial takes up an inordinate amount of screen real estate to show a very small data set. A typical dial shows no more than one or two data points as needles set against a background of two or three more points, representing acceptable and unacceptable values. This is typified by the example dial below representing a retailer’s inventory of a particular product, which I took from a Dundas Data Visualization dashboard demo. The dial itself plots only three data points: the current stock level (indicated by the pointer), the level below which a restock order should be made (the amber line) and the level deemed to be insufficient (the red line). The data density compared to conventional charts is very low and showing multiple such dials on a dashboard would create a cluttered, distracting effect.

Stock Gauge

Stock Gauge

2. The “pop out” value of the information on the dial is low, so the business message is not immediately obvious. To compensate for this in the above example, there is a separate light in the top left of the instrument to indicate a RAG status for the inventory value, which corresponds to whether the pointer is in the red, orange or white zones.

3. Representing a value as an angular orientation of a needle is not the most effective method of encoding data such that values can be perceived accurately. The research of William Cleveland and Robert McGill showed that human perception is more effective at determining accurate values if they are plotted as positional or length features in the chart rather than as angles.  This deficiency applies when gauges contain multiple pointers.

4. Reading the values directly from the scale can be very difficult. In the above example, the scaling algorithm in the dashboard software has scaled the instruments in major units of 200, each of which is divided into three minor units. This means that each subdivision represents a value of 66.67, which is hardly intuitive to interpret. To get around this problem, the actual values for the red and amber lines are shown beneath the dial and the value of the pointer has been made available on a tooltip. This again begs the question: ‘What insight is the dial adding?’

5. It is very hard to make comparisons between different parts of the business using dials, particularly when different scales are involved. The example charts below, from the same Dundas demo, illustrate this issue. The dashboard in question has a dropdown selector enabling the user to pick a geographic region. On switching from Europe to Pacific region, the scale of the instrument changes (from $0 – $600 to £0 – $500), which is not what the driver of a car would expect.

Europe Average Sale   Pacific Average Sale


6. Dials in a management dashboard lose the feature of the real car speedometer which shows at a glance whether the measure is increasing or decreasing and its rate of change. They are unable to show trends over time at all or to compare performance for different parts of the business on the same scale.

The bullet graph, developed by Stephen Few, is a much better alternative that addresses the key deficiencies of the dial or gauge.  It encodes the value of a single measure, enables a comparative to be displayed (e.g. budget or target), and displays it in the context of performance ranges, such as bad, satisfactory, and good.  Bullet charts possess a far stronger “pop out” effect than dials or gauges and can easily be stacked in horizontal or vertical groups to allow comparisons of several measures at once.

Bullet Graph

Bullet Graph

Leave a Reply