Course contentsShow
← Data Science
Lesson 1519 of 2,14537. A/B TestingFree lesson

Common Surrogate Metrics in A/B Testing

Review frequently used surrogates like click-through rate, engagement time, sign-ups, and page views with interpretation.

Common Surrogate Metrics in A/B Testing

You'll learn the most frequently used surrogate metrics in A/B testing and how to interpret them effectively.

Why We Use Common Surrogates

When measuring the impact of a product change, you often can't wait months to see if revenue increases or customer lifetime value improves. Instead, you measure surrogate metrics that respond quickly and predict those long-term business outcomes. Think of them as "leading indicators" — early signals that suggest where your business metrics are headed.

Frequently Used Surrogate Metrics

Click-Through Rate (CTR)

The proportion of users who click on a link, button, or feature. Easy to measure quickly and often predicts engagement and conversion downstream. Interpret cautiously: higher CTR doesn't always mean better user experience or revenue.

Engagement Time

How long users interact with your product or feature. More time can indicate value, but context matters — sometimes faster task completion is actually better (e.g., checkout flows).

Sign-Ups or Account Creations

A critical step in the user journey. High correlation with future revenue, making it a strong surrogate for acquisition-focused experiments. Watch for quality vs. quantity trade-offs.

Page Views or Sessions

Volume-based metrics showing overall activity. Useful for content or navigation changes, but can be misleading if users are confused and viewing more pages out of frustration rather than interest.

Interpretation Guidelines

Always validate that your surrogate moves in the same direction as your business metric. A 10% increase in page views means nothing if revenue stays flat or drops. Monitor multiple surrogates together to catch unintended consequences.

Key Takeaway: Click-through rate, engagement time, sign-ups, and page views are fast-responding surrogates commonly used in A/B tests, but each requires careful interpretation within the context of your actual business goals.