There's no hard technical cap on how many axes a radar chart can have — this tool will happily draw 20 if you add them — but readability breaks down well before any technical limit is reached. Here's where, and why, and what to do about it.
The practical bands
- 3 axes — the minimum; forms a triangle. Technically valid but visually closer to a simple comparison than what most people picture as a "radar chart."
- 4–8 axes — the sweet spot. Enough axes to show real multi-dimensional structure, few enough that every label stays legible and every point on the polygon is easy to trace back to its axis at a glance.
- 9–12 axes — usable, but labels start needing to shrink or wrap, and a chart with 3–4 overlapping datasets at this axis count starts looking visually busy. Still fine for a single dataset or two.
- 13+ axes — generally too many. Labels crowd into each other, adjacent axes become hard to tell apart at a glance, and the whole point of a radar chart — reading overall shape instantly — breaks down because there's simply too much to take in as one shape.
Why crowding happens specifically at the labels, not the polygon
The polygon itself can technically handle many points — it's just a connected line. The actual failure point is almost always the axis labels: with N axes evenly spaced around 360°, each one gets 360°/N of angular room, and text labels don't shrink proportionally to angular space the way the geometry does. At 6 axes each gets 60° of separation — plenty of room for a label. At 16 axes each gets 22.5° — labels for adjacent axes start overlapping or need aggressive truncation, especially the ones near the left and right sides of the circle where labels run more horizontally into each other.
What to do if you genuinely have more than 8–10 criteria
- Group into fewer, composite axes. Instead of 15 individual sub-metrics, combine related ones into 5–6 composite scores (average or weighted-sum the sub-metrics into one axis value per group). You lose granularity in the chart itself but gain readability — and can still present the underlying detail in a table alongside it.
- Split into multiple charts. Two clean 6-axis radar charts, each covering a different theme (say, "product specs" and "support quality"), are more readable than one crowded 12-axis chart trying to do both at once.
- Drop to fewer datasets instead. If the crowding is really coming from too many overlapping polygons rather than too many axes, comparing 2 items at a high axis count reads far more cleanly than comparing 4 items at the same axis count — each additional dataset adds visual clutter on top of whatever the axis count already costs you.
- Use a different chart type entirely. Past a certain point, a well-organized table or a set of small bar charts (one per criterion) genuinely communicates more clearly than a radar chart ever will at high axis counts — a radar chart is a deliberate tradeoff of precision for gestalt shape-reading, and that tradeoff stops paying off once there are too many shapes to read at once.
How this tool handles high axis counts
There's no artificial limit — add as many axis rows as you want, and the chart redraws live with even spacing regardless of count. That flexibility is intentional (some legitimate use cases genuinely need more axes than the "ideal" range), but the readability tradeoffs above don't go away just because the tool allows it.
- Is there a maximum number of axes this tool enforces?
- No hard limit — the geometry scales to any axis count you add. The readability guidance above is a design recommendation, not a tool restriction.
- What's the minimum?
- Three — fewer doesn't form a closed polygon shape.
- Does axis count affect how many datasets I can compare?
- No — the dataset limit (4) and axis count are independent settings; you can combine any axis count with 1–4 datasets, though the readability advice above compounds when both are pushed high at once.
→ Build a chart (free)