Make the
message
measurable.
We investigate the words between a good offer and a customer who understands it.
Explore the methodology ↗Small changes.
Sharper hypotheses.
Examine the words, the reasoning, and the numbers. These are illustrative reconstructions for learning—not claims about named clients or observed results.
Choose the plan that fits your business.
Choose the plan that fits your business a plan by the number of projects you run.
A concrete selection criterion may reduce the effort of comparing plans.
Before you test,
do the math.
Estimate the audience a two-variant experiment needs, or check the difference between two observed conversion rates. A useful result begins with a realistic question.
Sample-size calculator
Enter your baseline and the smallest relative lift worth detecting. The estimate uses a two-sided test, 95% confidence and 80% power, with equal allocation and independent visitors.
Significance calculator
Compare independent visitors and conversions in each group using a two-sided pooled proportion z-test. A p-value describes compatibility with equal rates; it is not the probability that a variant is best.
Friction is rarely
just a button.
Above the fold, a visitor is deciding whether the page deserves another second. If the headline asks them to decode a mechanism, infer an audience, and trust a promise all at once, attention can disappear before the offer gets a fair hearing.
Clara Vane begins with the reader’s next decision: what must be understood, believed, or resolved before a click makes sense? We turn that question into a hierarchy, then into a testable change. Scroll-depth and eye-tracking observations can reveal where attention shifts; they cannot tell us motivation on their own.
Meet Clara Vane ↗A rate is not
the whole result.
These hypothetical counts illustrate the difference between an absolute percentage-point change and a relative lift. They are not a recorded experiment. Use the calculator above to assess the example, and consider uncertainty before treating any observed difference as a decision.
| Cohort | Visitors | Conversions | Rate |
|---|---|---|---|
| Control | 2,000 | 80 | 4.0% |
| Treatment | 2,000 | 96 | 4.8% |
A change from 4.0% to 4.8% is an increase of 0.8 percentage points, or 20% relative to the original rate. Both descriptions matter: the first describes the practical change per visitor, while the second expresses its scale against the baseline.