Google Optimize is a simple, easy to use testing and personalization platform. Google has worked diligently to bring powerful AB and MVT testing tools to the market for both small businesses and large enterprises.
For small businesses, Google Optimize will offer free, easy to create AB and MVT testing that will be linked to their Google Analytics account with rich targeting options and Bayesian statistical analysis. As of now Google Optimize is the only major player with a completely free, no hit limit option (normally usage fees are incurred after an amount of monthly unique visitors or hits). There are three types of tests available:
AB Tests – AB tests are used to compare an experiment against the baseline. In Google Optimize these experiments can have up to 7 variations and 1 baseline.
MVT Tests – MVT tests isolate experiments into modular experiences which create a large number of total combinations via the rule of product. The free version of Optimize allows 4 sections and up to 16 total combinations. This is more than enough variations for most small to medium sized businesses since the more you slice your conversion volume, the less predictive power you will have.
Redirect Tests – Redirect tests are especially effective for single page tests where you have development resources creating assets to test rather than deploying content via the testing tool WYSIWYG editor.
When a test is concluded, Google Optimize will provide a winner alongside the range of improvement expected and the probability to beat the baseline experience.
In addition to top level metrics, you will see conversion rate trends so that yo u can visually confirm the Google Optimize statistical output.
During the setup for your Google Optimize test, you will integrate your tests with Google Analytics automatically. This will allow for deep analysis of how your website’s audience interacted with the test. You can use the “Variant” dimension to measure your metrics by device, channel, landing page, and several other secondary dimensions.
Convinced to create a test using Google Optimize? To begin your test we have created some easy to follow instructions.
This is the tricky part. Typically AB testing tags go above all other scripts. It is possible to deploy Google Optimize via Google Tag Manager, but it is not the best practice because it would delay firing anything until after gtm.js (the Google Tag Manager library) was loaded.
Google recommends a combined Analytic-Optimize script with page-hiding to reduce flicker (flicker is the visual phenomenon that occurs when a testing platform swaps out the original content for a variant).
Usually an AB/MVT test is triggered by a page or group of pages, but sometimes a website requires that a test fires when some other type of condition exists. Your developers can trigger tests by deploying an event to the data layer.
After your variants have been created, push the test live and let it run! Typically you want to make sure that you have a significant amount of conversion volume to begin with. As a rule of thumb I like to make sure that I can get at least 100 conversions for each variation over 6 weeks. If you cannot muster that number of conversions then maybe consider optimizing towards a micro transaction (like viewing cart instead of purchasing).