All Things Growth and Marketing

About two years ago, I wrote an article on using Google Tag Manager (GTM) to personalize your website. Even then, people asked why I wouldn’t just use Google Optimize. At the time, the answer was simple: Personalization was part of Google’s six-figure paid solution.

However, in November 2018, Google released the functionality to all users. Since then, Google Optimize has become a primary platform to initiate personalized experiences. But GTM is still critical to overcome its enduring limitations.

Keep reading

Facebook A/B Testing

Having a well-thought-out plan for A/B testing Facebook ad campaigns is essential if you want to improve your performance reliably and consistently.

And the more you test, the better. A study of 37,259 Facebook ads found that “most companies only have one ad, but the best had hundreds.”

A/B testing Facebook ad campaigns can get complicated quickly (and easily produce invalid results). Spending the time upfront to perfect your testing process and structure will go a long way.

Keep reading

Multivariate testing

Multivariate testing (MVT) is an experiment that tests changes to several elements of a page at the same time, such as the headline, image and call to action, to find out which combination performs best and how much each element contributes. An A/B test compares whole versions of a page; a multivariate test compares every combination of the elements you change. That makes it powerful and traffic-hungry: three headlines, two images and two buttons already make 12 combinations, each of which needs its own sample.

A/B testing vs multivariate testing vs split URL testing

A/B or A/B/n testMultivariate testSplit URL test
What changesOne or more whole versions of a pageSeveral elements at once, in every combinationTwo different URLs, often full redesigns
Question it answersWhich version wins?Which combination wins, and which element matters most?Does the new page beat the old one?
Traffic neededSample per version x number of versionsSample per combination x number of combinations, usually far moreSame as an A/B test
Best forMost tests, especially on low or moderate trafficHigh-traffic pages where you want to tune several elementsRedesigns and back-end changes
Main riskLearning less about why a version wonUnderpowered results and false winners across many combinationsRedirects and tracking errors

The traffic math decides most of the time. If a page needs about 2,000 visitors per variation to detect the effect you care about, an A/B test with two versions needs about 4,000 visitors. A 12-combination multivariate test needs about 24,000. Check your own numbers with the A/B test calculator before you choose. For the fundamentals, start with the A/B testing guide.

When should you use multivariate testing, and when is A/B/n testing best?

The answer is both simple and complex.

Keep reading

Categories

The marketing job paying $145k to $300k

It's called marketing engineer. US companies hiring for it ask to see what you've built. In CXL's live cohort, you build a working AI marketing system on your own data, and you're certified on that system, not on attendance.

Black Friday: a year of CXL for $1,599 $1,199. Includes the January cohort and every cohort after it.

Join for $1,199