S13 · WHICH TOTAL · FIGURE 07 The share of GDP is the number most likely to mislead — so it is built to be argued with. Turn the handle and watch where the peak goes.

FIGURE 07 Advertising as a share of GDP — the strip

How to read this

still mark = grade A — an audited / settled reading. (This strip's readings are grade B and C; each prints its grade under the mark.)
shimmering mark = grade C — uncertain; the amplitude is the uncertainty. Here the 1914 and 2025 endpoints are grade C and ride the edges.
cyan = an independent number — an audit, a census, a court finding.
amber = a seller-owned number. This figure is NEUTRAL: share of GDP is zinc scaffolding, so ownership rides only the source dots below — not the marks.
struck red = a corrected / DEAD claim. The popular "peak in 2000" (ds-gdp-001) was rejected in verification — it is drawn on the strip and stamped, never hidden.

Reduced motion: every shimmer becomes a static hatched band — denser for a softer grade — so the uncertainty is still seen, just not moving.

y-axis basis — advertising as % of US nominal GDP · non-removable caveat: the 2025 reading (e7-scale-002) is on a different basis from the 1919–2007 Coen/BEA series, and the 1914 reading is on the pre-1919 benchmark basis — a maximum is only a maximum inside one ruler.
Eleven dated readings of advertising's share of US GDP, drawn as marks and never joined into a line, because the record holds no annual series. On the post-1960 window 2000 is the highest reading on screen; widen the window to the whole record and 1922 (3.0%) rises above it. The rejected "peak in 2000" claim is drawn and stamped, not hidden.
Ce1-scale-009 Be2-scale-004 Be2-scale-005 Be2-scale-006 Be3-scale-005 Be3-scale-004 Be4-scale-003 Be5-scale-004 Be6-scale-003 Ce7-scale-002 Cds-gdp-001 · REJECTED
This figure could not render.