Claude Hopkins / our application for evaluation · Evaluation
Test the message in the real world
Claude Hopkins held that advertising exists to sell, so it should be judged by what it gets people to do, not by whether we like it. Be specific, test versions against traceable responses and separate what you believe from what the measurement actually shows.
When it helps
When you're choosing between headlines, offers or versions, or want to know whether a change actually helps the reader and the business.
Keep in mind
When the sample is too small to say anything, or the measurement mixes the copy up with changes in traffic, audience, season or price. The result is then a hint, not a verdict.
How to apply it
- Judge copy by what it gets the reader to do. Hopkins argued that advertising's job is to sell; whether copy is polished, witty or popular internally tells you nothing about that.
- Be specific. Hopkins observed that people believe exact statements but skim past generalities and superlatives. State what's included, what it costs and how it works, using facts you can stand behind.
- Make responses traceable. Hopkins used keyed coupons to see which ad brought replies; today that means separate links, codes or test groups for each version. Decide in advance what counts as an outcome.
- Change one thing at a time and write the hypothesis down: what you're changing, why, for whom and which outcome you expect.
- Report honestly. Your own judgement or a small sample is not a measured improvement, and a result from one channel or audience doesn't automatically apply to others. The modern measurement rules here are our application, not quotations from Hopkins.
Questions to work with
- What is the hypothesis, and what would show it's wrong?
- Which behaviour counts as the outcome – a click, an enquiry or a purchase?
- Can the measurement separate the copy's effect from other changes?
- Is there enough data for a conclusion, or is this only a hint?
A fictional example
Fictional: a cleaning company wants to test two headlines on its quote page. Neither has been tested yet.
Draft
The new headline will increase conversion.
Revised
Hypothesis: a headline that says what the price includes will bring in more quote requests. We'll show both versions at random and compare requests once we have enough data.
A promised effect becomes a testable hypothesis with a clear outcome – the results decide, not a hunch.
Read the original sources
The application and example are our own. The example is fictional and has not been conversion-tested.