How to Successfully Run Your First Website A/B Test
14 likes6,686 views
The document provides a comprehensive guide on how to effectively conduct A/B testing for websites, highlighting the significance of doing thorough research and understanding the data before implementing tests. It emphasizes the importance of formulating hypotheses, documenting results, and iterating on tests for continuous improvement. Additionally, it offers practical advice on what to test, the necessary traffic for significant results, and the tools that can aid in A/B testing.
How to Successfully Run Your First Website A/B Test
- 1. `How to Successfully Run Your First Website A/B Test THUE MADSEN
- 3. The don’ts… The dos… 1 What is A/B testing? Do your homework Repetition Succeed 2 Where do you start testing? TABLE OF CONTENTS 3 A framework for success Tooltips and Burgers4
- 4. What is A/B testing? A SCIENTIFIC METHOD OF PROOF
- 5. WHAT IS A/B TESTING FOR MARKETING • A scientific method adapted for marketing • Compare two or more versions of an asset to each other • Use data and statistics to determine which version is better • Implement winner and increase conversions
- 6. *Image from Optimizely HOW DOES A/B TESTING WORK?
- 7. *Image from Kikolani HOW DOES A/B TESTING WORK?
- 8. *Image from Dynamicyield HOW DOES A/B TESTING WORK?
- 9. SO WHY DO WE A/B TEST? • Lowers risk • Helps us realize the actual impact of proposed changes • Increase learnings • Takes the guess work out of optimization
- 10. Where do you start testing? THE DON’TS
- 11. *Image from Optimizely WHERE DO YOU START TESTING?
- 12. DON’T JUST GUESS
- 13. SUB OPTIMIZE • Button colors • Small copy changes • Small layout changes • Fonts, font size, etc. * at least not until you have found a great variation and have lots of traffic
- 14. WATCH OUT FOR BUZZ • Don’t just test what the gurus say • What works for others may not work for you
- 15. PREMATURE TESTING • Testing without traffic • Before you know what and why you are testing
- 16. Where do you start testing? THE DOS
- 17. DO YOUR HOMEWORK You have to put in the work to get great results • Learn as much as you can about A/B testing • Study the great people out there • What’s their process and insights?
- 18. A framework for success The basis of a successful testing program
- 19. Research Know what to test and see the greatest improvements
- 20. DIVE INTO YOUR DATA PaidEntered CCAdded to CartViewed Product
- 21. FIND YOUR BIGGEST OPPORTUNITIES PaidViewed CheckoutAdded to CartViewed Product
- 22. INVESTIGATE THE OPPORTUNITY TO UNDERSTAND IT And/Or
- 23. FORM A HYPOTHESIS By doing A, B will happen, Because of C
- 24. FORM A HYPOTHESIS By updating branding and increasing form visibility (A), We will get more prospects to register for webinars (B), Because prospect attention will be drawn to the form (C).
- 25. BUILD THE TEST
- 26. LAUNCH THE TEST VS. Control Variant
- 27. THE TEST VS. +24.01% increase in registrations!! 100% Significance
- 28. DOCUMENT AND SHARE YOUR FINDINGS
- 29. DOCUMENT AND SHARE YOUR FINDINGS Documenting all your tests and results • Increases company transparency • Helps others know where tests are going on • Makes it easier to call tests • Makes learnings available for everyone in company
- 30. *Image from Optimizely REPEAT THE PROCESS AS OFTEN AS POSSIBLE
- 31. A STORY OF OUR FAILURE VS Control Variant
- 32. A STORY OF OUR FAILURE VS
- 33. A STORY OF OUR FAILURE VS -13.68% increase in registrations!! 80% Significance
- 34. *Quote found on VWO NOT ALL TESTS ARE WINNERS ““Only 1 out of 8 A/B tests have driven significant change.” Noah Kagan AppSumo* ““The goal of a test is to get a learning, not a lift. With enough learnings, you can get the real lift,” Dr. Flint McGlaughlin MECLABS** **Quote found on MarketingSherpa
- 35. RULES OF THUMB FOR TESTING Things to remember when we are testing • Min. 2000 visitors • Min. a few hundred conversions • At least one full cycle > 2 week • >99% statistical significance • Keep a backlog of test hypothesis • Build next tests as the current one is running *More info at http://kiss.ly/1vF4wUc and http://kiss.ly/1Q4vTm0
- 37. How I like my testing burger Tooltips for your testing framework
- 38. Tool Company Why do I like it? Analytics Kissmetrics Shows me my actual customers and their behavior Surveys Qualaroo and/or SurveyMonkey Easy to use/Deep qualitative insights Project Management Trello Manage my time + communicate with collaborators
- 39. Tool Company Why do I like it? Company Wiki Confluence Document initial thoughts for everyone Testing platform* Optimizely, Pardot, Adwords Easily build tests and get data Analytics Kissmetrics Let’s me confirm and verify down my funnel Company Wiki Confluence Documentation and learnings for everyone *Includes actual testing platform and tools that have testing capabilities
- 40. Not everyone likes their burger the same way How do you take yours?
- 41. Trust the process Share your leanings Stack your results A/B TESTING FOR SUCCESS
- 42. KISSMETRICS IS ANALYTICS TO OPTIMIZE YOUR MARKETING www.kissmetrics.com
- 43. JUSTIN THORP Senior Marketing Operations Manager, AddThis @thorpus thorp@addthis.com THUE MADSEN Marketing Operations Manager, Kissmetrics @ThueLMadsen tmadsen@kissmetrics.com Questions?
- 44. Appsumo reveals its A/B testing secret: only 1 out of 8 tests produce results MORE RESOURCES… A Beginner’s Guide To A/B Testing: An Introduction How to Run AB Tests that Get REAL Results Statistical Significance Does Not Equal Validity (or Why You Get Imaginary Lifts) How Big is the Difference Between Being 90% and 99% Certain? My 7 Rules for A/B Testing That Triple Conversion Rates Optimizely: What is A/B Testing How To Determine The Statistical Significance Of An A/B Test – Basic Introduction