You should test that: How to use A/B testing in product design
Download as PPTX, PDF0 likes233 views
This document discusses user experience (UX) testing, specifically A/B testing. It provides an overview of A/B testing, including that it involves creating two versions of a page and splitting traffic between them to see how users interact. The document outlines why testing is important, different types of testing tools, considerations for running a successful A/B test such as defining goals and metrics, and examples of famous A/B tests. It stresses that testing takes time and one should not expect overnight results.
You should test that: How to use A/B testing in product design
- 1. YOU SHOULD TEST THAT! Kelley Howell, UX Product Manager
- 2. Why should we test? How we think users will act is often different than how they really act It is the best way to settle disputes about designs
- 3. Why should we test? Testing provides a direct way for us to identify users and their interactions with the website
- 4. Other tools to find out how we’re doing Usability Testing Field Studies Eye Tracking Click Mapping Multivariate Testing A/B Testing
- 5. What is A/B Testing?
- 6. A/B Testing is… • Two versions of a page are created • Traffic is split between versions • Users have a 50% chance of seeing one of the versions
- 7. It is important that only one element on a page is changed. Any more than that creates a multivariate test which has a different set of testing parameters.
- 8. Advantages of A/B Testing
- 9. Easy to create and implement tests
- 10. Flexibility in choosing testing objectives
- 11. Can be adapted for any traffic numbers
- 12. Measures actual behavior of users
- 13. Popular A/B Testing tools
- 15. How to A/B Test
- 16. Before you start testing • Prioritize pages • Define website goals • Create an hypothesis of what you expect to occur • Define which metrics you are going to measure and what those metrics are going to measured against
- 17. Decide which one thing you want to test
- 18. Know what variable you are testing for Determine which variable to test
- 19. Redacted Create alternatives Create Alternatives
- 20. Redacted Devote time to let results mature
- 21. Redacted Test, test, and keep testing!
- 22. What Can you A/B Test?
- 23. Call to Actions
- 24. Microcopy
- 25. Page Layout
- 26. Form Length
- 27. A/B testing can’t answer everything
- 28. Famous A/B Tests
- 29. Fact: A simple change was worth 300 Million Dollars
- 30. Fact: Google infamously tested 41
- 31. Fact: This A/B test made A/B testing cool
- 32. Things that can ruin your A/B test
- 33. Selecting Certain Population Sources
- 35. Stopping Your Test Early
- 36. Disadvantages in A/B Testing
- 37. 1) Results change as time marches forward
- 38. A/B testing: one at a time
- 39. At a certain level of maturity, dramatic results usually require dramatic design tests
- 40. 4) Not all A/B Testing Goals are testable
- 41. A/B Testing in a nutshell
- 42. Define your goals Make up a hypothesis to test Create a variation of your original Set goals on what to fine Run you test for long period Test, and keep testing to see if result holds true
- 43. That’s a Wrap! Remember: Testing takes time, don’t expect overnight miracles– and if that happens
- 44. For more research-based insights, check out the UX insights portal: http://redacted.com THOUGHTS? QUESTIONS?