Increase your online sales by 50% or more
We Improve Conversions and Key Metrics
We begin by using Omniture, WebSideStory or Google Analytics to measure your current web site traffic and performance metrics.
We then use the information we gather to identify critical areas to test and improve.
Finally, we provide statistically-significant, measurable improvements that show a clear and immediate ROI.
We Provide Insights About Your Customers
We help determine what technology your customers use, what search terms they use to find your web site, how many visits they make to your site before purchasing and much more.
Armed with this valuable profile, we then work with you to refine marketing messages and web site content and navigation.
We want to ensure that any changes we make to your web site can be measured and that improvements can be easily quantified.
To accomplish this, we use A/B testing. We split incoming traffic into two groups. The A group is the control group that continues to experience the original web site. The B group sees the new web site. The performance of each group is measured and changes are clearly seen.
A/B testing is effective, but it can be very slow when a variety of messages and ideas need to be tested.
Multivariate testing allows for the simultaneous testing of a variety of features on a web page such as wording, colors, images, and layout. Formulas are then applied to identify the optimal combination of the elements being tested.
This method helps us reach an optimal presentation for your web site in a fraction of the time that it would otherwise take.
Real-time Traffic Division
With both A/B testing and Multivariate testing, it is important to run the tests simultaneously to achieve reliable results.
If tests are run at different times (even back-to-back), the results cannot be trusted. Something as simple as a snowstorm in the midwest or a telecommunications problems in California could skew the results in favor of one test or another.
We employ modules in a variety of server-side scripting languages to randomly divide traffic in real time in order to run tests simultaneously and produce results you can trust.
Determining Statistical Significance
When making changes to a currently functioning web site, it is tempting to react too quickly to initial results.
In some cases, early returns on a new idea are higher than what the end result will be. Reacting too soon may lead to disappointment if expectations are set or actions taken based on the initial results. Similar problems arise when reacting too quickly to disappointing results.
We use statistical methods to show accurate, up-to-the-minute differences between tests along with a margin of error. We help you know when results are statistically significant so you can make informed decisions.
Analyze Traffic and Establish a Baseline
When applying web analytics to a client’s web site, we prefer to begin by analyzing current traffic to establish a baseline.
This step is needed to better understand customers visiting the web site and to measure improvements that we later make.
With the data we collect in step one, we form hypotheses about how we can improve conversions and key metrics on the web site.
We then implement the ideas and prepare them for testing.
Once our hypotheses are ready for testing, we deploy them in either an A/B testing environment or a Multivariate testing environment.
Statistically significant results are measured and appropriate action is taken. The return on investment for the effort is clearly visible. The process can then be repeated again to further improve results.
HomeVirgo needed to update their web site, but they were nervous to do so because of a negative experience they previously had when an “upgrade” turned out to be a “downgrade”.
They had many features on their web site that they felt provided great value to their customers. Intuitively, most of the features seemed to make sense.
However, they really didn’t know what was providing real value and what was distracting visitors and keeping them from making a purchase.
HomeVirgo decided to have their web site redesigned before applying extensive analytics. The only baseline measurement taken was subscription signups per day.
Our redesign of the web site maintained the majority of the original features arranged in a new, user-friendly manner. In depth analytics were added as a part of the new web site.
After the deployment of the new site, we analyzed their traffic in detail. Using this valuable information, we made some additional enhancements to the web site and revamped the primary marketing message on the site.
The results were astounding. The initial redesign of the web site doubled the amount of signups each day.
The enhancements made after gaining knowledge of their customers added an additional 70% improvement over their baseline results.
In the end, HomeVirgo nearly tripled their signups per day using our web analytics approach.