I’m always happy to hear when clients want to expand their website testing programs and include other areas like paid search SEM. There are four common ways people test SEM combined with Adobe’s Test&Target testing platform.
Scenario 1 – Important first step is to set up a reporting ‘segment’
This first recommended scenario may not be as intuitive as the typical test scenario where all SEM traffic gets its own A/B/n test. It is important to first answer the question: ‘Do SEM visitors behave differently than the rest of the visitors?’ If so, then in what ways are they different?
To gather this data, first create an A/B/n test to all of your visitors, including paid search visitors. Then set up a reporting segment for your paid search SEM traffic. Once the test runs you can see the winning test experience for your overall population and you can compare it to the performance of the SEM traffic. If one test experience/variation wins for you overall visitors and a different experience/variation wins for your SEM visitors, then you now have data showing that these two groups behave differently.
One variation of this scenario is to create multiple SEM reporting segments inside the T&T interface. For example, set up a reporting segment for all branded keywords, set up a segment for each product/category, and then a segment for all SEM.
Any page with T&T mbox code on it will automatically read in the URL text of that page. Create a reporting segment inside T&T and set it to track visitors when their landing page URL contains the SEM URL tracking parameter (e.g. “trackingId=”, “SEM=brandedKeywords”, “cid=”, or some other parameter that is unique to all paid search).
Scenario 2 – A test targeted only to SEM traffic
This is the typical test scenario where the A/B/n test is created to only be shown to SEM paid search traffic. All that is needed is the T&T mbox code on any page of your site where SEM traffic will be landing. Simply set up a test campaign and target the campaign to only show to traffic when the visitor lands with a particular URL parameter. As mentioned above, this is typically the tracking ID of SEM campaigns but it can also be any text string inside the URL.
The test content above can be shown on just the landing page or across the entire site. For returning visitors, you can choose whether to show or not show the test content again.
Scenario 3 – A test that excludes SEM traffic
Sometimes the SEM testing and general testing needs to be separated. This is easily done by repeating the steps above but this time target the test campaign when the URL does not contain the URL parameter.
Scenario 4 – SEM ad text is being A/B/n tested and you want the website content to match the SEM ad text seen.
In this scenario, the A/B/n testing is at the ad text level and Test&Target is not A/B/n rotating the content. Instead, that rotation is controlled by your search provider. T&T is showing the content that corresponds to the SEM ad text. For example, if you are testing the SEM ad text ‘Free 2 week subscription’ vs. ’10% off monthly subscription’, then the landing page and other site pages can show content corresponding to those two versions.
The easiest way to implement this is to add a different URL parameter to the end of each ad text’s destination URL. (For example, www.yoursite.com?NormalURLparametersHere&SEMtest=A vs. www.yoursite.com?NormalURLparametersHere&SEMtest=B.) With this approach you can then target each experience in T&T with the corresponding unique parameter.
All this functionality to create testing rules around the URL is all built in and fairly easy to set up inside Test&Target. Additionally, T&T offers specific keyword targeting if you don’t want to use the URL parameter. Furthermore, you can also expand this concept beyond paid search to other advertising channels.
Paid Search is typically a very important part of a company’s marketing program and a significant portion of the marketing budget is going into it. Testing and optimizing this program will help you to get the most ROI from it.