This entry is part of the Chart Index, the reference library for the Chart Design Field Guide.
The gauge is the dashboard chart everyone has used and almost no expert recommends. Modelled on physical instrument dials — speedometers, fuel gauges, pressure meters — the gauge encodes a single value as the sweep of an arc within a dial. The form is intuitive (everyone has used a car's instrument panel), and that intuition is what keeps it in dashboards despite well-documented perceptual shortcomings.
Stephen Few has written extensively against the gauge. His argument: a gauge wastes 60–80% of its pixels on chrome — the dial face, the needle hub, the bezel — and encodes a single number as angle, one of the least accurate perceptual channels. A bullet chart conveys the same information in a fraction of the space. This article exists to document the form, set its limitations explicitly, and point readers toward better alternatives.
What it is
A gauge maps a single quantitative variable to the angle of a needle or the sweep of a filled arc within a circular dial (typically a semicircle, but full-circle and quarter-circle variants exist). The dial is divided into qualitative bands — good, acceptable, poor — traditionally colour-coded with traffic-light schemes. The needle or arc indicates the current value; the dial's start and end mark the minimum and maximum.
The hero chart above is the house gauge: a customer satisfaction score of 78 on a 0–100 dial, drawn as a filled arc over a light track rather than a needle, with the value printed large at the centre. The thin ring outside the arc carries three qualitative bands — poor, fair, good — and the black tick at 85 is the target. The reader can see that the score is in the middle band and short of the target; what they cannot do is judge how short without the number. Randomise it and the arc sweeps to the new value, clamped to the dial.
When to use it
Gauges are the right choice when:
- The audience expects the form and would find a bullet chart unfamiliar (industrial control rooms, automotive contexts, classic instrumentation dashboards).
- The visual metaphor of an instrument dial carries domain meaning (engineering, manufacturing).
- A single number with qualitative context (good / bad bands) is the only message.
- Screen real estate is not constrained and you want a focal point — kiosk displays, control rooms.
These conditions are narrow. Most dashboard contexts are better served by bullets, sparklines, or labelled values.
When not to use it
- Multi-metric dashboards. Twelve gauges in a grid is twelve circles fighting for attention. Use bullets.
- Comparing values across dimensions. Gauges show one value; comparing them across panels requires the eye to estimate angles in three different gauges. Use bars or bullets.
- Precise reading. Angle perception is imprecise; readers can identify roughly where the needle is, but not precisely. For precision, pair with a numeric label.
- Trends over time. Gauges are static; they show the current value with no history. Pair with a sparkline or use a sparkline alone.
Design principles
Use semicircular, not full-circle
A semicircular gauge uses the visual space efficiently and has clear start (left) and end (right) anchors. Full-circle gauges introduce a discontinuity at the top (12 o'clock) and put the start and end of the scale in the same place, so a value of 0 and a value of 100 point at the same spot.
Pick band thresholds carefully
The qualitative band thresholds — good above 85, acceptable 70–85, poor below 70 — are the chart's interpretive content. They are domain decisions, not visual choices, and they should be documented. Without thresholds, the gauge is just an angle.
Use desaturated bands
Saturated traffic-light colours (red / amber / green) draw the eye to the background bands rather than to the data. Use quiet greys or pale tints for the bands so the value arc is the only strong colour on the dial — the house gauge steps three greys around the outside for exactly this reason.
Prefer the bullet for the same information
The gauge and the bullet below carry the same four facts: the value, the target, the bands and the scale. The bullet does it in a strip a fifth of the height, and — because it encodes the value as length on a common baseline — the reader can judge the shortfall against the target far more accurately. Add two more KPIs and the bullet simply grows two rows; the gauge would need two more dials.
Show the value as text
A gauge without a numeric value annotation is incomplete. Place the current value prominently — typically below or inside the arc. The angle gives the rough position; the number gives precision.
Show the target as a tick mark
If the gauge has a target, indicate it with a small perpendicular line across the arc. The reader compares the arc's end with the target tick. Without the tick, the bands are the only context for whether the value is good.
Mark the scale endpoints
The minimum and maximum should be stated — 0 at the left, 100 at the right, or in the caption. Without scale labels, the arc's sweep is uncalibrated: a dial that runs 0–100 and a dial that runs 50–100 look identical at three quarters full.
Pair with a sparkline or trend ribbon
A gauge alone is a snapshot. Paired with a small sparkline showing the last quarter of values, the gauge becomes more useful: where we are now with where we have been. The sparkline below is what the hero gauge is missing — the score has been sliding for six weeks, which no dial can show.
Anatomy
A gauge's anatomy is a semicircular track, a value arc, qualitative bands, a target tick, scale endpoints, and a numeric value. The form is dense for a single value and uneconomical for multiple values.
Related types
- Bullet chart — the modern replacement. Same information, less space, more precise.
- Progress bar — for actual against 100%, without bands.
- Sparkline — for trend rather than position.
- Numeric KPI card — just the number, often with a delta. Often more effective than a gauge.
- Radial bar — several gauges' worth of progress as concentric arcs; shares the gauge's angle encoding and its weaknesses.
Reading list
- Few, S. (2006). Information Dashboard Design. The original critique of gauges and the case for bullets.
- Few, S. (2013). Information Dashboard Design, second edition. The bullet graph specification.
- Tufte, E. (1990). Envisioning Information. On data-ink ratio and the cost of chrome.