Scaling your
Testing Program
Vivek Vilatia
Senior Product Manager, VWO
Anshul Gurbaxani
Associate Product Manager, VWO
Prem Kujur
Customer Success Manager, VWO
Ask Questions on Twitter
#AskVWO
What is A/B testing
In the most simplest form, A/B
Testing is comparing two versions
of a web page to see which version
performs better for a defined
conversion goal
Type of Tests
A/B Test
Change any element or section
of a webpage and test both
versions.
Split URL Test
Test performances of separate
URLs if the landing pages are
significantly different.
Multivariate Test
Run tests on multiple variation
combinations together. Needs
sufficient traffic.
Graduating to
Advanced tests
Simple A/B tests Advanced A/B tests
➔ Great starting point
➔ Smaller teams
➔ Easy to setup
➔ Quick wins
➔ Test changes like copy,
arrangement of elements,
CTA colors etc
➔ Test advanced use cases with
greater impact
➔ Conversion optimization program
➔ Self-sufficient team
➔ The right tool-set
➔ Management buy-in
➔ Tons of perseverance
Audience Changes MetricsPages
Key steps in setting up your tests
Test pages | Defining a set of similar pages for your test
Asterix * in the URL matches pattern would replace the variable part of the URL
Represented by * in the URL matches pattern
Audience | Complex targeting rules
Visitors logged in on
your website
Are either from India or have their
preferred currency as INR
Changes | Code editor for complex changes
Changes | Code editor for complex changes
Changes | Conditional Test Activation
Changes | Conditional Test Activation
SignUp Form Type Form
Measuring the
right metrics
Measuring the right metrics
Micro Conversions Macro Conversions
Measuring the right metrics
Never make your tests
live without proper
sanity checks
Quick previews
Cross Browser testing
Live previews
Check cookies using a plugin
Making sense of your
A/B testing reports
Making sense of your A/B testing reports
Run your test for at least a week.
Along with a primary goal, have related secondary goals.
Ensure that your primary goal represents the direct and
exclusive impact of the changes being tested.
Making sense of your A/B testing reports - Best Practices
Slice and Dice your test results to learn the trends and
behaviour of various user segments.
Testing radical changes at times might just do the trick
Identify micro-conversions and important correlative metrics
increase in a given variation, implement it.
Inconclusive Tests
Ensure that you don’t use segments with very less data.
Make sure there is no error in the test setup.
Ensure that you are not running multiple tests
simultaneously in the same part of the journey you are A/B
testing
Inconclusive Tests - Pitfalls
Thank You!
Q&A

VWO Webinar: Scaling Your Testing Program

  • 1.
    Scaling your Testing Program VivekVilatia Senior Product Manager, VWO Anshul Gurbaxani Associate Product Manager, VWO Prem Kujur Customer Success Manager, VWO
  • 2.
    Ask Questions onTwitter #AskVWO
  • 3.
    What is A/Btesting In the most simplest form, A/B Testing is comparing two versions of a web page to see which version performs better for a defined conversion goal
  • 4.
    Type of Tests A/BTest Change any element or section of a webpage and test both versions. Split URL Test Test performances of separate URLs if the landing pages are significantly different. Multivariate Test Run tests on multiple variation combinations together. Needs sufficient traffic.
  • 5.
  • 6.
    Simple A/B testsAdvanced A/B tests ➔ Great starting point ➔ Smaller teams ➔ Easy to setup ➔ Quick wins ➔ Test changes like copy, arrangement of elements, CTA colors etc ➔ Test advanced use cases with greater impact ➔ Conversion optimization program ➔ Self-sufficient team ➔ The right tool-set ➔ Management buy-in ➔ Tons of perseverance
  • 7.
    Audience Changes MetricsPages Keysteps in setting up your tests
  • 8.
    Test pages |Defining a set of similar pages for your test Asterix * in the URL matches pattern would replace the variable part of the URL Represented by * in the URL matches pattern
  • 9.
    Audience | Complextargeting rules Visitors logged in on your website Are either from India or have their preferred currency as INR
  • 10.
    Changes | Codeeditor for complex changes
  • 11.
    Changes | Codeeditor for complex changes
  • 12.
    Changes | ConditionalTest Activation
  • 13.
    Changes | ConditionalTest Activation SignUp Form Type Form
  • 14.
  • 15.
    Measuring the rightmetrics Micro Conversions Macro Conversions
  • 16.
  • 17.
    Never make yourtests live without proper sanity checks
  • 18.
  • 19.
  • 20.
  • 21.
  • 22.
    Making sense ofyour A/B testing reports
  • 23.
    Making sense ofyour A/B testing reports
  • 24.
    Run your testfor at least a week. Along with a primary goal, have related secondary goals. Ensure that your primary goal represents the direct and exclusive impact of the changes being tested. Making sense of your A/B testing reports - Best Practices
  • 25.
    Slice and Diceyour test results to learn the trends and behaviour of various user segments. Testing radical changes at times might just do the trick Identify micro-conversions and important correlative metrics increase in a given variation, implement it. Inconclusive Tests
  • 26.
    Ensure that youdon’t use segments with very less data. Make sure there is no error in the test setup. Ensure that you are not running multiple tests simultaneously in the same part of the journey you are A/B testing Inconclusive Tests - Pitfalls
  • 27.