# Lina's Channel ROMI — Schema

Flat CSV in `lina-channel-romi.csv`. One row per channel.

## Columns

| Column | Type | Description |
| --- | --- | --- |
| `channel` | text | Marketing channel name (`Paid Social`, `Email`, `Search`, `Content`, `Display`). |
| `leads` | integer | Total leads generated by the channel in the reporting window. |
| `cost_per_lead_usd` | integer | Blended cost per lead in USD. |
| `conversion_rate_pct` | decimal | Lead-to-customer conversion rate as a percentage. |
| `romi_ratio` | decimal | Return on marketing investment as a ratio (revenue attributed / spend). Values above `1.0` are profitable; below `1.0` lose money. |

## Row inventory

- 5 rows — one per channel.

Total: 5 rows.

## Deliberate design choices

The dataset carries all four metrics per channel so the reader can reproduce Lina's original clustered-column chart (which plots leads, cost per lead, and conversion rate across three incompatible axes) as well as the single-encoding rebuild (which uses only `romi_ratio`). Keeping all four columns visible makes the contrast tactile: the reader can see that only one of the four metrics answers the VP's Q4-spend question, and that all four columns being present is exactly why the clustered column fails.

`romi_ratio` is quoted to two decimals to match the Big Idea's exact wording (`$3.40`, `$0.80`). Any downstream chart that displays these values matches the print figure exactly.

## License

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