Optimizely

Chris Castiglione @castig

OPTIMIZELY: A/B TESTING MADE EASY

VOCABULARY OPTIMIZELY: A/B TESTING MADE EASY

OPTIMIZING LANDING PAGES FOR CONVERSION

VOCABULARY A/B Test

Call to Action (CTA)

Experiment

Control (or “The Original)

Multivariate Test

WHY A/B TESTING?OPTIMIZELY: A/B TESTING MADE EASY

A/B TEST

HYPOTHESIS #1: LONGER PAGE

HYPOTHESIS #1: LONGER PAGE

HYPOTHESIS #2: PERSON

HYPOTHESIS #3: LONG PAGE + PERSON

HYPOTHESIS #3: LONG PAGE + PERSON

HYPOTHESIS #4: CHANGING PEOPLE

HYPOTHESIS #4: CHANGING PEOPLE

OPTIMIZING LANDING PAGES FOR CONVERSION

CONCLUSIONPage length is an important variable

Big photos of smiling customers work

SETTING UP AN OPTIMIZELY EXPERIMENT

OPTIMIZELY: A/B TESTING MADE EASY

IMPLEMENTING THEPROJECT CODE

OPTIMIZELY: A/B TESTING MADE EASY

Q&A

Q: What should I test?

WHAT TO TEST Short page vs. Long Page

Headline Text

Buttons

Form fields

Q: Is it possible to target specific users?

TARGETINGLanguage

Browser

Visitor Type (new vs. returning)

IP

Referrer

Custom Events

Q: Can I specify the amount of traffic that comes to my experiment?

SPECIFY THE AMOUNT OF TRAFFIC OPTIONS -> Traffic Allocation

Q: What effect does the JavaScript project code have on page performance?

PAGE PERFORMANCE Around 35-40kb (very small!)

The code run from Optimizely’s CDN