6 Steps for Building a Well-Oiled Content A/B Testing Machine

Step 1: Document everything

  • Learning from successful tests and attempting to replicate similar successes
  • Learning from failed tests and making sure not to repeat them
  • Tracing patterns that can’t be identified as a result of a single test
  • Promoting alignment, transparency, and collective learning within the organization
  • Serial number
  • Landing page
  • Test type
  • Current status
  • Start date
  • End date
  • Hypothesis
  • Results per KPI
  • Statistical significance level

Step 2: Classify A/B tests by “type”

Step 3: Formulate strong hypotheses

Step 4: Research your target audience

Step 5: Prioritize. Prioritize. Prioritize.

  • While testing on desktop, focus on the area above the fold
  • Opt for test types that tend to generate positive results and high monetary value
  • Run tests that have proven successful on other similar landing pages
  • Avoid tests that have failed across multiple similar landing pages

Step 6: Regularly sync and share knowledge

Now go ahead and make mistakes

--

--

--

UX Writing Team Lead — Wix.com

Love podcasts or audiobooks? Learn on the go with our new app.

Recommended from Medium

When I heard word of its closure, my first thought

How to Learn Data Science For free : just do it!

About Record Linkage and the “Golden Record”

Is your Data telling the story?

My Tryst with Deep Learning — German Traffic data set with Keras

Twitter Sentiment Analysis Using ULMFiT

Interested in Data Science?

#USAID on twitter : a quick NLP analysis

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Bar Zukerman

Bar Zukerman

UX Writing Team Lead — Wix.com

More from Medium

Finding Yourself — Personal Value Proposition

Expanding Access Through Digital Therapeutics?

How Demographic Factors Influence Consumer Behavior

How observing agents in action led to a feature that saves 30 seconds of handle time.