Building KPI Micro Charts

Look at any dashboard people actually use daily and you'll find the same thing across the top: a row of small, dense indicators. A number, a tiny trend line, a gauge. No axes, no titles, no legends — just the state of the business in a strip you can read in two seconds.

That row is made of ordinary widgets with most of their decoration turned off. This lesson builds it.

First: clone the dashboard

Before you start stripping settings off widgets, make a copy. CRM Analytics doesn't give you a deep undo history, and this kind of work is inherently trial and error.

The timeline chart for date trends

The first KPI is usually a trend: revenue over time, cases over time. The instinct is to reach for a bar chart grouped by month. Reach for the timeline chart instead.

The difference is that a timeline understands that the axis is time. A bar chart treats your month values as just another set of categories — equally spaced, in whatever order the grouping produced, with gaps for empty periods silently collapsed. A timeline places points on a genuine time axis, spaces them proportionally, and handles missing periods honestly.

For a trend, the line is also just the better shape: the eye reads direction and slope from a line far faster than it compares the heights of forty bars.

Stripping it down to a micro chart

Now make it small. A KPI tile is maybe two inches wide, and at that size every piece of chart furniture is noise. In the widget properties, turn off:

• The x-axis and y-axis, including tick labels and axis titles
• The chart title and subtitle — the tile's own label already says what it is
• The legend, which is meaningless for a single series
• Gridlines and any background fill

What survives is the line itself: pure shape, no scaffolding. That's a micro chart — the same idea as a spreadsheet sparkline. It doesn't tell you the exact value at any point, and that's fine, because the number sitting next to it does. The micro chart's only job is to answer "which way is this going?"

The projected line

Timeline charts have a feature bar charts can't offer: a projected or predictive line. In the timeline's reference-line options you can enable a forecast segment that extends the trend forward, usually rendered dashed or lighter so it's visually distinct from actuals.

There's one requirement people trip over constantly: your data has to span today's date. The projection extends from the present into the future, so if your dataset ends last quarter — or your dashboard filter cuts it off before today — there's nothing for the projection to anchor to and the line simply doesn't appear. It looks like the feature is broken. It's the filter.

The polar gauge

For a metric with a target, a gauge beats a number. The polar gauge renders as a partial ring that fills toward a maximum you define — attainment against quota, cases closed against SLA, adoption against goal.

Its strength is that it shows position within a range at a glance: three quarters of the way round is instantly legible as "nearly there" in a way that "76.4%" isn't. Set the maximum to something meaningful (the target itself, so a full ring means goal met) and apply conditional formatting so the fill colour also carries the status.

Its weakness is space: a gauge takes several times the pixels of the number it represents. Use one, maybe two, for the metrics that genuinely have targets — not for everything.

The number widget

The workhorse of the KPI row. A single measure, rendered large. Two settings make it dashboard-ready:

Between them you get a KPI that communicates magnitude and status with no chart at all — the cheapest widget on the dashboard in both pixels and query cost.

Assembling the row

Line up your tiles across the top: number plus micro chart for trending metrics, a gauge for the one with a hard target, plain formatted numbers for the rest. Give the whole strip a consistent treatment — the dark widget theme from the previous lesson works beautifully here — and drop the cell spacing so it reads as one band rather than five boxes.

Everything below that band is detail. The band itself should answer the question "is everything OK?" without scrolling.

Your dashboard now has a headline. The last piece is connecting it to everything else — next up: linking dashboards.