references/canon/hopkins.md

# Hopkins

## Principle

Prefer measurable response, specific reasons to believe, and controlled testing over advertising decisions based only on taste.

## Use when

- Offers, headlines, landing pages, mail, paid acquisition, or promotions can be measured.
- The team is debating preferences without behavioral evidence.
- A claim or offer can be tested at modest cost.

## Do not over-apply

- Not every brand effect is immediately attributable.
- Short-term response metrics can undercount future or indirect value.

## Decision questions

- What response can be observed?
- What variable is actually being tested?
- What result would change the next decision?
- Is the test measuring business value or only a proxy?

## Counterweight

Pair response testing with longer-horizon brand and incrementality thinking when effects are delayed or indirect.

## Sources

- `source:hopkins-scientific`