were going to be looking at one of Google Analytics most recent additions to its feature set; Experiments. Using this tool, Ill be showing you how to serve up different variations of a page to determine which one is the most successful in converting visitors to the site.

If youve ever created a website, youll almost certainly be familiar with Google Analytics. With a free Google account and a snippet of code, you are able to track visitors to your sites and drill down on the data with startling https://www.tradebit.comm small personal projects to enterprise level sites, Google Analytics has established itself as the market leader for very good reasons; its free, simple to implement and is suitable for the casual user or even the most battle-hardened marketer.

Ready to get started? Lets begin!

By serving up different versions of a page to simultaneous visitors, were effectively doing what people in marketing circles refer to as Split Testing. The ultimate goal of this approach is to determine which of a group of webpages (or brochures, banner ads, text ads, etc.) converts visitors most effectively. While a conversion could be just about anything, typically website owners are interested in purchases, signups, account registrations or even getting visitors to follow them on social channels.
The most important concept to grasp about split testing is that the variations must be run all at the same time. In other words, it would not be considered an effective split test to load up one page for a month, collect the data and then upload a different page to collect and compare the data again. A key part of split testing is to ensure that as many factors as possible that could influence a site visitor are removed.

Up until fairly recently, A/B Split testing (i.e. testing between two variations) was the standard approach for testing online conversions. The idea with A/B split testing is to arrive at a winner and then rinse and repeat several more times, each time getting closer to the page that converts the maximum amount of visitors. While fairly effective, A/B Split Testing has some critical limitations which well discuss in a moment.