# Marcus's Compromise Dashboard — Chapter 2 Worked Example

> The two-audience compromise-dashboard scenario from Chapter 2 of *The Defensible Decision*. Six visuals on one page that fail both audiences, alongside the two-BRD alternative: one small chart per audience, each answering its audience's specific decision question.

**Wholly synthetic.** Any resemblance to a specific real sales-ops dashboard is coincidental; the pattern is drawn from the recurring practitioner failure mode Chapter 2 catalogues.

## Files

- [`marcus-compromise.csv`](marcus-compromise.csv) — flat rows: six compromise-page visuals + two two-BRD solution rows
- [`marcus-compromise.json`](marcus-compromise.json) — canonical scenario contract with source hash

## What it represents

The Page 1 Marcus would build if he tried to serve the VP of Cloud Sales Operations and the regional directors from one merged BRD. The compromise has six technically-correct visuals (a card, a line, three bars, a table, a treemap) that together fail both audiences: the VP cannot identify three defund candidates from the page, the regional director cannot identify two products to push in NA-East. The two-BRD alternative pairs one purpose-built chart per audience, each answering its audience's specific question in a defined read-time.

## Invariants (locked)

The scenario carries executable invariants that Chapter 2's release contract enforces:

- Compromise dashboard: exactly six visuals, IDs `1..6` with no gaps.
- Visual kinds span at least four distinct chart families (card, line, bar, table, treemap) — the failure mode is not a single-encoding-choice bug.
- Neither audience's decision question is answerable from the compromise page (`vp_outcome` and `director_outcome` both record the recovery step the audience is forced to take).
- Two-BRD solution has exactly two charts, one per audience.
- Each solution chart names its audience, its decision, and its chart-shape rationale.
- Read-time budgets are specified per solution chart (`90` seconds for the VP chart, `150` seconds for the director chart), tying the audience discipline back to Chapter 5's read-time framing.

## Why the invariants matter

Chapter 2's pedagogical point is that a merged BRD produces a merged chart that fails both audiences. If the compromise dashboard had fewer visuals, it would look like a rushed job rather than a diligent one; if the two-BRD alternative used identical chart shapes, the comparison would collapse to a re-titling exercise. The scenario is deliberately shaped so the compromise looks like the reasonable-effort dashboard a mid-career analyst would ship, and the two-BRD path looks like the smaller, cheaper, more defensible alternative that pays back the up-front elicitation cost.

## What it does NOT represent

- Actual sales-operations dashboard data from any real company.
- Marcus's Page 2 or later pages (the failure is visible on Page 1; the chapter does not extend to a full dashboard tour).
- The BRD templates themselves (those live in `/appendix-e/` — this dataset only carries the chart-shape decisions the two BRDs would drive).

## Reproducibility

The dataset ships from the book source at `AlexBooks/applied-ai/ai-directed-viz/data/marcus-compromise.json` and is re-emitted here by `AlexBooks/applied-ai/ai-directed-viz/scripts/generate-marcus-compromise.mjs --site-root <site-root>`. The generator writes the source SHA-256 into `marcus-compromise.json` so drift between the book and this site is detectable.

## License

Released for teaching use alongside *The Defensible Decision*. Use freely.