SIGNAL — Finding-First Data Visualization

Summary

Two series, two different units, eight quarters. The request implies one chart with both — which is the setup for a dual y-axis. I'm not going to build that, and the reason is arithmetic rather than aesthetic. Below: the question your data can actually answer, the finding, and the encoding that proves it honestly.

Data read — 8 quarters, 2 series. Subscribers in thousands (count). Ad revenue in $K (currency). Different units, non-commensurable. Both strictly increasing, no gaps, no outliers.

Analytical question

The request says "visualize this." The question underneath it, for a board audience, is:

Is monetization keeping pace with audience growth?

That's a question about relative rates, not absolute levels. It decides everything downstream.

Finding

Subscribers nearly doubled while ad revenue grew just 18%.

The arithmetic

236 ÷ 120 = 1.967. 484 ÷ 410 = 1.181. Audience grew more than five times faster than the revenue it generates.

The derived figure, if you want it in the notes: revenue per subscriber fell from

$3.42 to $2.05, a 40% decline. Same data, stated as a rate.

Recommended visualization

Both series indexed to Q1 2024 = 100, on one shared scale, with the band between them shaded.

Indexing is earned here, not imposed. "Keeping pace" is a ratio question, and indexing is the encoding that answers ratio questions directly. If you'd asked "what's our revenue per subscriber," the right answer would be a single derived series — not an index. The transformation follows the question.

The gap at Q4 '25 is

78.6 index points, and that gap is the finding — so the accent goes on the shaded band between the lines, not on either line.

Why the default chart is weaker

The default here is subscriber bars on a left axis running 100–250, revenue as a line on a right axis running 400–500. Both ranges are plausible. Every plotted value is correct.

And yet

A series that grew 96.7% and a series that grew 18.0% ascend the page by within

3.3 percentage points of each other. A board member reads "revenue is keeping up." That conclusion is manufactured entirely by four numbers the analyst chose, and nothing on the chart signals a choice was made.

A dual-axis chart has four free parameters. Any crossing, any convergence, any "these move together" can be produced from true values. That's why this one has no exceptions.

(The left axis is also a truncated bar axis — a second, independent defect. The dual axis is the one that matters.)

Design directions

Accent the band, not the lines. Fill the area between the two indexed series at ~10% opacity in a single accent color. The divergence becomes the subject.

• Direct-label both series at their right ends. No legend.

Two emphasized values, by exception: +97% in ink and +18% in grey, both large. The finding is the contrast between two rates, so both carry emphasis — separated by color tier, never by size. Setting one larger implies a precision ranking the data doesn't contain.

One annotation: 79-point gap by Q4 2025, placed inside the band's widest region. No leader line needed — the text is already standing on its subject.

• Everything else neutral. No chart border, no vertical gridlines, no markers on all sixteen points.

Integrity & accessibility

Transformation disclosed in the subtitle, not a footnote: Indexed to Q1 2024 = 100 · subscribers vs. ad revenue. An undisclosed transformation is the same class of failure as an undisclosed truncated axis.

• Shared zero-anchored scale. Both lines start at 100 by construction; the axis includes it.

Meaning survives grayscale — the two lines differ in weight and are directly labeled, so hue carries nothing on its own.

Alt text: "Line chart of subscribers and ad revenue, both indexed to Q1 2024 = 100, across eight quarters. Subscribers rise to 196.7 while ad revenue reaches 118.0; the widening band between them reaches 79 index points by Q4 2025."

• Source line: Source: internal reporting, Q1 2024 – Q4 2025 · indexed to Q1 2024 = 100 · illustrative data

10-point audit

10/10. Last check: would a different but equally defensible axis choice change the reader's conclusion? No — there is one scale, and both series sit on it.

Next step

Tell me your plotting environment and I'll write it — matplotlib, Plotly, Vega-Lite, ggplot2, or hand-authored SVG. If you also want the absolute levels for the appendix, that's two stacked panels sharing one x-axis, never a second y-axis.