Product6 min read

Conversion Rate Optimization

Systematically improve the percentage of visitors who take a desired action - through hypothesis-driven A/B tests, landing page optimization, and funnel analysis.

CROLanding PageA/B TestingConversionGrowth

skill: Conversion Rate Optimization category: Product tags: [CRO, Landing Page, A/B Testing, Conversion, Growth] readingTime: 6

CRO is a systematic process, not a bag of 'best practices'. Best practices are starting hypotheses, not guaranteed wins - your user base may be different. Run a proper A/B test before shipping any CRO change. The most impactful CRO is fixing a broken user experience, not button color.

Workflow

Identify conversion opportunities with dataCRO effort is wasted on pages with minimal traffic or where conversion is already high. Focus on high-traffic, high-value pages with measurable drop-off.
Write a testable hypothesisA hypothesis that predicts the mechanism of improvement is testable and teaches you something even when the test fails.
Optimize the landing page above the foldThe hero section (visible without scrolling) is the highest-impact area - 80% of users never scroll past it on first visit.
Optimize forms for completionEvery additional form field reduces completion rate by ~3-5%. Forms are the bottleneck between intent and conversion.
Design statistically valid A/B testsRunning an underpowered test or stopping early because one variant is 'winning' produces false conclusions - wasted effort and potential regression.

Output Format

CRO system: opportunity analysis, hypothesis backlog, A/B tests, and statistical analysis of results.

CRO Backlog:
  Opportunity: pricing page, 40% drop, $50K revenue potential
  Hypothesis: social proof will lift conversion 15%
  Test: 2-week A/B, 8000/variant, primary metric: trial start
  Result: +18% lift, p=0.02, shipped

Iterations: tested → winner → new baseline → repeat

Examples

Input

Output

Edge Cases & Gotchas

Novelty effect - A new variant wins in week 1 because it's different - by week 3 the lift disappears. Run tests for at least 2 full weeks before concluding.
Seasonal traffic effects - Running a test during Black Friday or end-of-quarter skews results toward atypical buyer intent. Exclude abnormal traffic periods from tests.
Multiple testing correction - Testing 10 variants simultaneously has 40% probability of a false positive at p=0.05. Use Bonferroni correction or test sequentially.
Conversion vs. quality of conversions - A test that doubles signups but halves paid conversion rate is a loss, not a win. Always measure downstream metrics (activation, payment) alongside the primary conversion metric.

When to Break These Rules

Fix broken UX before CRO - a confusing checkout flow needs to be redesigned, not A/B tested button by button.
Focus on traffic before CRO when your sample sizes are < 1000 visitors/month - tests will never reach significance and effort is better spent on acquisition.

Reference Files

https://www.cxl.com/blog/
https://www.evanmiller.org/ab-testing/sample-size.html