Leveraging Coremetrics Real Estate Tags to Perform A/B Testing

Dorian D. Regester — Sat, 09 Jan 2010 20:41:11 +0000

Real estate tags can be leveraged to track not only creative performance on a page but also A/B and multivariate testing. As you may recall, the real estate tag track how similar links on the same page perform. However, you can also leverage it to track A/B and MVT testing.Since real estate tags allow users to break down reports by creative version, creative position and link name, one can take the tag and pass in the components of the A/B test. Then you can see how the version is performing (A or B), which call to action (if there are multiple ones on the page) drove the user and .To do this you will need to tag every destination link on the A and B copies with cm_re parameters. The tag would look similar to:http://www.xyz.com/?cm_re=Holiday Sale Version A-_-holidaybanner_jpg-_-728x90Using the versioning in the first parameter, you can know exactly what copy is being tested without having to stuff too much into the parameter. Holidaybanner_jpg is the name of the image that drove the click. Putting the jpg or gif or fla can let you see which one drives higher conversion on your site (static vs dynamic). The 728×90 is the image dimensions and can be handy if you test many pages (then you can easily see my creative dimension which size is driving more people in. You might initially think its your larger images but in fact it might be a text link.However if you want to know more about the link (i.e. revenue and other conversion KPIs), you should also add a cm_sp or site promotion tag to it. Site promotion tags follow the structure of promotion group, promotion name and link (creative) name. If you wish, you can add in whatever parameters in place of the three out of the box ones (however if you are using site promotion tags elsewhere, you will want to follow that established naming convention).So the new tag would look like:http://www.xyz.com/?cm_re=Holiday Sale Version A-_-holidaybanner_jpg-_-728×90&cm_sp=Marketing-_-Holiday Sale-_-holidaybanner_jpg_728x90If you plan on using for multivariate testing, you could shape the tag similar to above just making sure each test case can dynamically pass in a new version value to the real estate tag.This is a nice initial way of A/B and MVT testing but by no means is it the best. This will help you build up the business case to bring in a more robust package to automate this process.

How To Run A/B Test in Omniture SiteCatalyst

Dorian D. Regester — Tue, 08 Dec 2009 23:23:27 +0000

Are you, like many, who have Omniture SiteCatalyst in-house but cannot afford to layer on Test&Target or another A/B testing solution? Well, welcome to the party lol…but seriously, just because you can’t purchase another solution for your company, does that mean you are left out in the cold? The simple answer to that is NO. This posting will cover the topic of leveraging SiteCatalyst to run A/B test and to take that new data point and correlate/subrelate against other variables.To leverage SiteCatalyst for A/B testing (synchronous), you can use a combination of sprops and eVars, which are perfect for A/B testing because they allow you to set a variable on a page and then wait to see what traffic or conversion event occurs. If you are testing two versions of a page, you can simply set the variables to a value that tells something about the given page, within either s_code.js or in the base code, when the visitors load the page. You can also use a traffic variable, called s.prop, in A/B testing for traffic and pathing-related data. If you are interested in knowing how different versions affect traffic and pathing on your site, then all you need to do is copy the value that you populated the eVar with into the sProp*.* Please remember that pathing is not automatically turned on for all sProps so you will need to contact Omniture ClientCare and have them enable pathing for that specific sprop.Below is a proposed solution for testing two different creative versions via Omniture:Let’s say you have two versions of creative and you want to see which one performs better. This is one of the simpler A/B test. In this specific case, all you need to do is assign a value to corresponding s.prop and s.evar on both the control and test page. Here is a snippet of code that could be used to accomplish this.s.pageName=”Home”s.prop1=”Control Version”s.prop2=”Control Version:A”s.eVar1 = “Control Version”Setting up the code like this would enable you to not only see pathing by creative version but also traffic reports (page views, unique visitors, avg time spent, etc) and commerce (orders, revenue, items, etc) broken down by those two new variables. Leveraging the pathing will require you to either set up the getAndPersistValue plugin or create logic that captures the creative version on the initial page and appends on the page name then on subsequent pages just passes the page name. On the other creative version, the following snippet of code would go.s.pageName=”Home”s.prop1=”Test Version 1”s.prop2=”Test Version 1:A”s.eVar1 = “Test Version 1”There are other forms of testing (segmentation and consecutive) that will be discussed at a later time.

Web Analytics Central » A/B & Multivariate Testing

Leveraging Coremetrics Real Estate Tags to Perform A/B Testing

How To Run A/B Test in Omniture SiteCatalyst