Analyzing landing pages and e-commerce transactions in Google Analytics

Posted by Josh on Wednesday, July 16th, 2008.

Suppose you have several landing pages for your AdWords campaign or other advertising program, and you want to know which such page is most effective at generating e-commerce transactions. You have Google Analytics reliably tracking e-commerce transactions. You won’t find this information in the e-commerce section, but there are several ways for Google Analytics to measure landing page effectiveness.

Goals
You can setup goals in your Google Analytics profile, one for each landing page. This will work, but you’ll quickly run out of goals, if you haven’t already. It also won’t work retroactively, thus you’ll have to wait to collect enough new data.

Website Optimizer
The Website Optimizer can help you do multivariate testing of landing pages and other content. Unfortunately it’s quite complicated relative the simplicity of our question. Like goals, this method does not work with old data either.

Hidden in Google Analytics
The data you seek is actually hidden in Google Analytics already, although not quite as conveniently as we might like. Go into your profile, select Content / Top Content. Click on the page you are interested in knowing about (you might use the filters / search box on the bottom to help you find the correct pages). Change the “Segment” drop-down box (below the graph) to “keyword”, “source”, or “campaign”. Now you should be able to click on the “Ecommerce” tab and see how many transactions were recorded that started with that page. Unfortunately this does not allow you to look at several pages side-by-side, but it’s easier than the other two methods and works on your past data. There is one BIG caveat to this data. Conversions are not simply those that started on this page, they include everyone that visited the page at any time during the session in which the transaction was completed. If the only way to get to the page is from an ad, then it’s a pretty safe assumption that it only acted as a landing page.

Thanks to Nic A.