Gauge & Speedometer Charts in Excel: When They Work (and When They Lie)

You came here to build a gauge chart in Excel, so here's the first thing to know: Excel doesn't have one. There's no gauge or speedometer in the chart menu — most dials you've admired are a doughnut chart and a pie chart stacked together with a formula pushing the needle. I'll show you how that's done, because now and then it's the right call. But the more useful thing I can tell you is when that dial earns its space and when it quietly lies to you — because the person who designed the chart built to replace it worked out exactly where the line is.

I build dashboards for a living — the reporting our managers open every morning at a multi-branch clinic — and I package one as a template. Two things follow. I'm not selling you an alternative to Excel, so I can be straight about it, including talking you out of the chart you came here for in the cases where it would cost you. And every design claim below is pinned to a named source at the end, because "gauges are bad" is only worth saying if you can check who says so, and why.

There's no gauge chart in Excel — you fake it

Open the chart gallery and look: column, bar, line, pie, area, scatter, and the newer additions like treemap and waterfall. No gauge, no dial, no speedometer. Microsoft has never shipped one as a chart type, in any version — while its sibling, Power BI, does include a native gauge visual. So every Excel gauge is a workaround: chart types that do exist, tricked into looking like a dial.

The most common method layers two charts:

• A doughnut chart for the dial. You size its slices to your coloured zones — say a grey stretch, an amber stretch and a red one — and hide the bottom half so the ring reads as a 180° arc instead of a full circle.
• A pie chart for the needle, sitting on top. It's built from three numbers: your value, a very thin sliver that becomes the visible needle, and whatever's left to complete the circle. The two big slices are made transparent, so all you see is the sliver pointing at the dial.
• A helper cell that turns your metric into those three numbers, so the needle swings when the data changes.

That's the usual recipe, and there are variants: some builds skip the pie and use a second doughnut for the needle, some drop the needle entirely and just print the number in the middle of the dial. All of them are fiddly — a lot of no-fill formatting, a bit of rotation, and a formula in the middle whose logic you'll have forgotten by next quarter. None of it is native, so none of it survives a careless edit especially well. If you want the click-by-click, Chandoo has the doughnut-and-pie-needle build and Excel Campus a needle-free two-doughnut variant (refined by Jon Peltier). I'm not going to reproduce forty steps here, because the steps aren't the interesting part. The question worth your time is whether you should.

When a gauge actually earns its space

Stephen Few is a large part of the reason "don't use gauges" is conventional wisdom among people who design dashboards for a living — he wrote the book on the subject (Information Dashboard Design) and returned to the point often. But he doesn't say never. As he put it in a 2010 essay, when we stray from the principles "we should do so because the situation demands it, which is at times the case. If we never learn the principles and understand how and why they work, we'll never know when it's appropriate to stray." So — when does the situation demand a dial?

One case holds up. A single, familiar KPI, measured against a target, shown to people who don't read charts for a living — a number on a break-room wall screen, a status light someone glances at between patients. The car-speedometer metaphor is doing real work there: "needle in the red" is widely understood, at a glance, without a legend — provided that red zone marks a threshold you can actually defend, not one picked to make the needle sit pretty (more on that below). If that's your whole situation — one number, one glance, an audience that would tune out a bar chart — a gauge is defensible.

Two honest caveats, though. That's the narrowest possible case, and it happens to be the exact job the bullet graph was built to do in less space (more on that below). And the "everyone understands a speedometer" argument is intuition, not evidence — it's plausible, I believe it, but I couldn't find a study that measures it, so I won't dress it up as proven.

When it lies — the five ways

Outside that one case, a gauge costs you. Here's how — and, where the point is someone else's, who made it.

1. It spends the room of a table to show one number

This is Few's bluntest objection: "Circular gauges — the darlings of dashboards — waste space, but there is none to waste on a dashboard where a great deal of information must be displayed in a single screen." A dial, its bezel, its tick ring, its coloured arc and its needle add up to a lot of pixels, and "by their very nature, circular gauges use a great deal of space to say relatively little." In Edward Tufte's terms it's a low data-ink ratio — most of the ink is decoration rather than data. One number, a whole card's worth of chart.

2. The needle is a hard way to read a number

A gauge asks you to read a value off the angle of a needle on a curved scale. A bar asks you to read it off a length on a straight one. Those aren't equal. In the foundational study of how accurately people read charts, Cleveland and McGill ranked the visual cues, and position along a common scale came out on top — the most accurately decoded of all; angle and area sat clearly below it. A bar uses the cue that wins; a gauge uses ones that don't. You can print the number in the middle of the dial to compensate — and you should — but then the dial is just scenery around a figure you could have shown plainly.

3. You can't compare them

This is the one that matters most on a real dashboard, because a dashboard is almost never one number. Put five gauges in a row and the reader has to take in each dial on its own, one at a time — there's no shared scale for the eye to run along. Few is unusually pointed here: gauges "fail spectacularly when intended for comparison." Line the same five metrics up as bars or bullet graphs on a shared axis and the comparison happens in a single sweep of the eye. Five dials is five reads; five strips is one.

4. The colour bands are an opinion wearing a lab coat

This one is my argument, not a citation — but it's the trap I see most. Most gauges carry coloured zones: a good stretch, a warning stretch, a bad one. The bands look like a fact about the world; they're almost always a judgement someone made. "Red below 90%" is a decision about what counts as bad, and the dial presents it with the authority of a measured instrument — the same sleight of hand this blog keeps flagging, a target dressed up as a measurement. A band is fine when it marks a real, defensible threshold — a contractual floor, a clinical limit; it's a problem when it's a number someone picked so the needle would land in the green, and the chart makes the two look identical. In fairness, this isn't unique to dials: the bullet graph I'm about to recommend carries the same coloured bands, and they need defending just as much. The failing is the invented threshold, not the shape that shows it.

5. It shows now, and nothing else

A dial is a snapshot. It tells you where the needle sits this second and says nothing about how it got there or where it's heading — and "we're at 82%" reads very differently if you're climbing than if you're sliding. Few concedes this is hard for any single indicator: even with a bullet graph, "it's not always so easy to tell if you are on track to meet or surpass that future target." The difference is that a bar-with-a-target sits happily beside a sparkline, and a dial leaves no room for one. If the trajectory is the story, the gauge is the wrong container.

The chart that does the gauge's job in a strip

There's a specific answer to "what should I use instead," and it isn't a matter of taste. Stephen Few designed a chart for this exact job — one measure, its target, a good/warning/bad range — and he designed it expressly to retire the dial. It's the bullet graph, which in his words "was developed to replace the meters and gauges that are often used on dashboards," with a linear design that "not only gives it a small footprint, but also supports more efficient reading than radial meters."

It's the actual-vs-target bar, refined: a horizontal bar for the value, a marker for the target, and two or three quiet background bands for context — all in a strip about the height of a line of text. Everything the gauge was trying to say, in a fraction of the room, on the cue the eye reads best. And because it's linear, the strips stack, so you compare many at a glance — the very thing the gauge can't do. Few made the contrast concrete when he lined the two up: the bullet graphs showed the same variables "in less space and in a way that makes comparisons much easier and faster." Their edge, he wrote, "is due in large part to their linear design."

There's even a small experiment on it. An eye-tracking study by Wu and colleagues, in a 2018 Springer proceedings, put the two forms in front of forty people and found the bullet graph was read more efficiently than the gauge. One modest study, not a verdict — but it points the same way as the theory.

And here's the practical part for an Excel user: you don't need the doughnut-and-pie contortions at all. A plain actual-vs-target bar — the value bar plus a marker for the goal — is genuinely less work than a gauge and reads better. A full bullet graph, with its graded context bands, takes a little more to build than that bare bar, but still nothing like the needle rig. Either is layout three in the Excel dashboard examples, and drops into the standard interactive build.

If you still want to build one

Sometimes you'll build the gauge anyway — the boss likes it, the break-room screen calls for it, the situation genuinely demands a dial. Fair enough. Build it well:

• Keep it to one KPI. The moment you want a second dial to compare against the first, you've walked into the case gauges fail at. Switch to bars.
• Always print the exact number. The needle is decoration; the figure is the information. Don't make someone estimate a value off an angle when you can just show it.
• Set the colour bands from a target you can defend, and say what it is. If "red" means "below our 90% contractual floor," label it. If it just means "a number that felt bad," you're colouring in an opinion.
• Don't build a wall of them. A grid of gauges is a strong sign a dashboard was made to impress rather than to inform.

Frequently asked questions

Written by

Olha · clinic data analyst

I build the reporting our managers open every morning at a multi-branch medical clinic — and package it so other practices don't have to start from scratch.

The design claims here are attributed to their original authors and quoted only where I could confirm the exact words at the source: Stephen Few on gauges and the bullet graph, from his 2010 Perceptual Edge article and his Bullet Graph Design Specification — both read directly, not from second-hand paraphrases, which is how a line Few never wrote once ended up attributed to him across the web. Cleveland and McGill are cited for how accurately we read different chart cues; Tufte for the data-ink ratio, which I apply to gauges myself, since he never wrote about them. I checked and rejected a "gauges are 31% faster, per ISO 9241-210" claim that turns up in AI-generated summaries: it traces to no primary source, and ISO 9241-210 is a design-process standard that publishes no such figure. The "everyone understands a speedometer" argument is presented as intuition because I found no study that measures it. The illustration is original, not a screenshot of anyone's product. Lucid Vitals is not affiliated with Microsoft.