|
Split Testing
Split testing allows you to effectively test
two different variables, and then measure the
effect that the differences have on a customer's
behavior. Minor changes in website content can
often have a significant impact on the number
of completed sales.
Many webmasters and marketers fail to spend the
necessary time thoroughly testing their websites,
and they often make assumptions about what looks
the best, what will perform the best, etc. Unfortunately,
those assumptions are not always correct. The
only way to truly know how different variables
will impact a website visitor's behavior is to
conduct a split test.
The concept behind a split test is not terribly
complex. In most cases of split testing, half
of the visitors will see one version of a given
page, while the other half will see an alternate
version. Webmasters track the actions taken by
the visitors when presented with each version
of the page or site, and then analyze the results.
When the traffic volume is significant, meaningful
conclusions can be drawn regarding which version
of the page was most effective in guiding the
visitors to the desired action.
There are a variety of things that software developers
can split test and analyze. The items a MicroISV
may want to consider testing include...
Promotion And Marketing Content
Changing the format or copy of any given promotion
can impact the close rate for sales.
Buy Now Vs Trial
Some developers may want to try offering one webpage
that offers a downloadable trial version, and
another webpage with only a "Buy" option, to see
how it impacts sales.
Product Reviews Or Testimonials
Adding product reviews to a website may influence
a visitors behavior.
Money-Back Guarantee
Including a money-back guarantee on the purchase
page may increase the sales close rate.
"Buy Now" Button Text
The actual text contained on the purchase button
can also be tested. Using "Buy Now" vs "Order
Now" may influence the order close rates.
"Buy Now" Button Design And Placement
The size, color, and placement of the "Buy Now"
button can also impact sales.
Box Shot Vs Screen Shot, And Placement
Providing a physical box shot or a program screenshot,
as well as the placement of those images on the
webpage, can all be tested to determine what is
most effective in increasing sales.
Search Engine Landing Pages
The landing pages for Pay-Per-Click marketing
campaigns can be tested to determine which version
results in the most visitors taking the desire
action.
Home Page Banners
Advertising banners on the home page can be rotated
to determine which banners are most effective
and generate the highest number of click-throughs.
Tools For Split Testing
Hi
Conversion
Split
Tester
Split
Test Accelerator
Split
Test Calculator
Keep in mind that the sampling of visitors must
be large enough so that the statistics are not
unfairly skewed. The larger the sampling, the
more accurate the results of the split test will
be.
Related Articles:
How
Accurate are a Software Developer's Statistics?
Referring Sites
Statistics
Emetrix
Sales Tracking Trends
Days
to Sale
|
Developer
101
Effective Split Testing