Small multiples compare shapes, not precise values

What it is

What small multiples produce are shape-grade comparative conclusions: "this group's curve runs steeper," "this panel's bimodality is unique," "the anomalies cluster in the lower-right." They do not produce value-grade conclusions: per-panel resolution is limited, and reading an exact value from one panel is both difficult and unreliable. Using small multiples as a readout instrument — "read Q3's value off the third panel" — mismatches the tool; the correct division of labour has small multiples discovering (which panel deserves a look) and drill-down views or tables verifying (what that panel's number actually is).

Why it happens

Read-off depends on the scale-annotation-alignment trio being locally available at the target position: within a small panel, ticks are sparse, annotations trimmed, grids thin and dense, and interpolation error from data point to reference grows as the panel shrinks. Shape comparison tolerates low resolution — trend direction, peak-and-valley placement, and overall silhouette survive downsampling, exactly the information small multiples preserve best. The reading division follows: the eyes do pattern matching across the array, while numeric queries go to interactive readouts (hover, click-through) or a companion summary table.

Where it stops holding

Key values may be labelled directly on a few panels (peaks, latest), giving shape conclusions a numeric anchor — but labels consume in-panel layout too, usually capped at two or three. Interactive hover readouts join "seeing shapes" and "reading values" seamlessly: the array discovers, hover verifies. Tasks needing batch precise read-off (reconciliation, report figures) should have been tables from the start, not charts.

Applying it

• Confine small-multiple conclusion sentences to shape language (steeper / flatter / bimodal / cliff), annotating anchors or switching to drill-down views when exact values appear.
• Pair the array with hover readouts and a click-through detail layer; attach a summary table beside the report version.
• Verification: ask readers to read a specific value from a panel and compare with truth; large error is not the reader's fault but the medium's limit — route it through annotation or tables.

Related

• Same group: U2.15.1 Small multiples replace overlay with repeated like-for-like panels · U2.15.2 All facets must share the same scale to be comparable · U2.15.3 Facet order should carry information, not alphabetical habit · U2.15.4 Too many facets shrink panels below readability
• Nearby: U2.16.1 Tables beat charts when exact values are needed · U6.09 Hover details and touch alternatives
• Search terms: shape comparison · readout · drill-down · small multiples