A/B (A/B Test)
A/B is shorthand for A/B testing — a controlled experiment where two variants of a page, ad, email, or feature are shown to randomly assigned users to determine which performs better.
What it is
An A/B test (also called a split test) randomly splits traffic between two or more versions of an experience. Each variant's conversion rate is measured. With enough sample size, statistical analysis declares a winner with a stated confidence level (typically 95%).
Why it matters
A/B testing replaces opinion-based decisions with evidence. "We think the red button will work better" is a hypothesis; an A/B test is proof. Without it, design and copy decisions are based on intuition — and most intuition-led changes have no effect or hurt.
Common A/B tests for healthcare landing pages
- Headline copy (benefit-led vs problem-led)
- Hero image vs hero video
- Phone CTA vs form CTA above the fold
- Short form vs long form
- Insurance badges vs accreditation badges
- Trust testimonial placement
Frequently asked questions
How long should I run an A/B test?
Until you reach statistical significance with sufficient sample size — typically 2–4 weeks for paid-traffic landing pages. Stopping early is the #1 way to fool yourself with false-positive 'wins'.
Can I test more than two variants?
Yes — A/B/C or multivariate. Multivariate needs significantly more traffic; only valuable when you have lots.
What's a meaningful lift?
Depends on baseline. A 5% relative lift on a 2% baseline is statistically demanding; a 30% relative lift is much easier to detect. Pursue bigger swings on low-traffic pages.
