According to Tim Ash, President of SiteTuners.com, online marketing can be split up into three essential categories. 1) Acquisition (getting people to the website), 2) Conversion (getting them to do what you want on your website), and 3) Retention (getting them to become repeat visitors and/or customers). Ash feels that most marketers are focused on acquisition and retention, but they should be paying more attention to conversion. "The part that's really been missing is what happens when somebody hits your site," he says. When a Web consumer first visits your website, they arrive at what is commonly referred to as a "landing page. As Ash sees it, "That's where the new battleground is."
To maximize the efficiency of landing pages, testing is needed. For example, if you are testing three headlines, four offers, and six button colors, then there are seventy-two (3x4x6=72) possible versions or "recipes" in your test. Below are three types of testing available:
- A/B Split Testing (1-10 recipes) - Good for head-to-head testing of one variable at a time. Easy to set up.
- Taguchi / Multivariate Testing (10-100 recipes) - Can test a small number of variables simultaneously. Tricky to properly design test and analyze the results.
- Website Tuning (1,000,000+ recipes) - Designed for large-scale tests. Takes variable interactions into account. Requires high data rates.
A/B Splits and Taguchi testing cannot scale up to large test sizes and also suffer from a more fundamental flaw - they assume that all elements being tested are independent of each other. Website tuning allows testing of millions of variables and design combinations in relation to each other. An example of this is SiteTuners.com TuningEngine. "As online marketers, we want designs that create synergy," says Ash. "We're designing for interactions.
In the end, they are able to pick out the best performing one. The results are then verified in a head-to-head test against the client's original site.
SiteTuners also has a unique business model. It is 100% performance based - taking a percentage of the new site's increased profits. If they can't improve your website, they don't get paid. They are also very selective about their clients, focusing on midsize and large online businesses with significant online traffic.
So does it work? SiteTuners claims conversion improvements from 20-200%, with a median of 50-60%. "We can change the economics of their online program overnight," says Ash.
So what about the smaller Web-based business? Ash recommends regular testing, of any kind. "You can do it in-house and cheaply rent A/B split testing software to get started. Testing is something that you should be doing with every mission-critical change that you are considering to your website.
And for all sites, he echoes a common adage, "Less is more."
SiteTuners.com was launched on May 1, 2004, in response to a need. Epic Sky, SiteTuners parent company, runs performance-based PPC campaigns. Their clients and affiliate merchants received plenty of traffic, but few conversions - because the sites were not tuned properly.