Your creative strategy will be influenced by your overall program conversion objectives. You may be starting out with a new website and wish to optimize all of the creative before you think about any testing. You may have extensive statistics on hits, clicks and sales or other conversion performance…with a wish to improve your performance and/or make it more cost efficient. You will work with our experienced staff to develop a strategy that makes sense for your business.

Either way, it is important to have a solid strategy. There are three basic methods of testing which companies use to improve performance. They are:

Split testing is the most basic and typically only requires three to four sets of creative. With this method,you are trying to see how your control performs against three or four new creative alternatives. These alternatives with either contain intuitively developed alternatives that are comprised of different sets of headlines, sub-headlines, graphic elements, body copy and layout (templates), you will work with our strategists and creative staff to develop three to four distinctively different packages that key into your product or service.

A/B split testing is typically the most economical and the choice of many companies beginning with conversion optimization testing.

The multivariante technique is perhaps the most sophisticated and very likely the most expensive. It is, however, statistically project able, extremely scientific and great for sites with complex product offerings. You can test any type of web page in any combination. You create a series of variables for the test that can include: landing pages, registration pages, pop-ups, credit card pages, and list building pages.

Each variable is typically plugged into one or more statistical models that predict success. The most common and often most effective model uses a statistical technique, Regression Analysis. Typically you start with Linear Regression. which assumes a logical data distribution and has proven to be an excellent predictor in the direct response environment. The next step would be a Logistic Regression Model which incorporates probability analysis. This technique has been used successfully by many direct marketers.

Multiivariant testing typically requires 10-15 complete sets of creative.

This method uses a smaller number of creative elements used in specific combinations to reach similar conclusions to that of multivariate testing without the need for large volumes of traffic. This method can also save time over multivariate testing on the same elements.