GPWA Times Magazine - Issue 24 - April 2013

This page is your first step in creat- ing generations of tests and producing useful data. Think of your control page as a bird and variations as eggs that have been laid. The egg that hatches, or the winning variation, be- comes the next generation’s control to “lay more eggs,” or be used for future variation testing. Creating a good struc- ture will make all the difference in your testing success, because it will make each change less time consuming. One of the reasons split testing is so attractive to mar- keters is that it is the least compli- cated method of market testing and is generally more fruitful in its results. How to implement With a good, strong control page as your beginning test subject, you are now free to brainstorm, research and build theo- ries and ideas regarding what you believe will be most effective in gaining conversions. Live by the KISS (keep it simple, stupid) phi- losophy and maintain only one hypothesis per variation as this allows for more concise testing and results. The idea is to find the vary- ing element on your page that converts the best and then build from there. There are several im- portant elements of your page that are available to be tested that when utilized effectively can re- ally drive home that conversion rate. They include: Headline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Gives visitor the first impression of your brand Success is measured by how closely it matched what the viewers were expecting to see when they clicked through to your page People are generally impatient; they will read your headline quickly and move on, so make it count Images . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Color vs. black & white Pictures of people vs. pictures of product One image vs. many images Male-oriented images vs. female-oriented images Copy and text . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Positive vs. negative language (“save time” vs. “stop wasting time”) Focus on the benefits first and then corroborate with facts and function Social proof (testimonials, endorsements, etc.) Video . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Auto play vs. press play Calls to action Length of video Image vs. video Hired spokesperson vs. staff spokesperson Number of videos Lo-fi vs. Hi-fi Forms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Number of fields Relevance of fields in relation to your offer Design (what style box) Positioning on page (left, middle, right) Title of form (call to action*) Privacy policy link (improves trust) Button text *An important tip to note is that when constructing your “call to action” aim to finish the sentence “I want to . . .” (e.g., “read more,” “get offer” or “enter now”). How to measure results Now that you have a list of possible hypotheses, you can con- duct your split testing accordingly. It is a good idea to be very detailed with your testing. Keep notes and records in relation to all processes that occur as this will aid you im- mensely in future testing and in the analysis of the data you produce. Depending on the amount of traffic that you have “By interpreting your data correctly and seeing that there was a statistical significance bet- ween A and B you can move forward know- ing you’re making an informed decision.” 52 Split Testing