Discover how W3i – an application distribution company that connects publishers, advertisers, and customers – used multivariate testing to achieve a conversion rate lift of 162% for Profile Pimp for MySpace software.
What happens when you combine the strength of Google Website Optimizer Multivariate testing capabilities with InstallIQ’s proprietary Experimentation Platform? That’s what we wanted to know.
One of our primary focuses in the W3i Marketing team is translating creativity into big revenue gains for our partners. In January 2008, we released the first version our proprietary Experimentation Platform that allowed us to test the user experience inside our installation manager, InstallIQ, which is the engine that drives the W3i Application Network. Over the next year we tested things as small as skins and as big as a fully customized experience based on demographics and split by traffic sources, often with double digit gains. However, as we tested more, it became harder to achieve big lifts.
Earlier this year I began hearing the same issues from our partners. Not only were they struggling to achieve bigger gains, many also faced tighter resources, shifting user behavior and technology challenges. We decided to assemble an internal team and go back into the test kitchen. We had dabbled in Multivariate testing a couple of times, but it was expensive and we weren’t able to find a partner that could tie into our Experimentation Platform to test all the way through to revenue. Which, let’s face it, is where we all need to be.
About the time the team was ready to hit the streets to start gathering intelligence, we received a call from our friends at Google. They were working on their “Google Jumpstart” program and were wondering if we did any Multivariate testing? Timing is everything! Shortly thereafter, we met the amazing folks at WiderFunnel Marketing Optimization, a Website Optimizer Authorized Consultant, that doesn’t just “consult” but provides all the services required to design and execute a conversion rate optimization strategy.
Both our teams worked closely over several weeks and the results of our combined efforts are truly amazing. We are incredibly excited to be able to offer our partners stellar conversion rate improvements in these challenging economic times. Kristen Groom, Sr. Marketing Specialist at W3i, has written the case study below detailing this experiment. I hope you find the possibilities it holds as exciting as we do!
Enjoy!
Lisa Nistler
Sr. Director of Marketing
W3i, LLC
W3i MARKETING CASE STUDY: Increase Landing Page Performance by 162% using Google Website Optimizer
I can hear the sighs now – another article promising to increase your landing page performance. While it’s true and feasible to achieve our success on your landing page, the extent will likely vary depending on your niche and level of existing optimization.
There are a few killer take-aways from our experience so read on to find out what you can do to maximize your landing page performance while taking advantage of the Google Website Optimizer.
WiderFunnel Marketing Optimization
W3i recently had the opportunity to work with a company called WiderFunnel, a Conversion Optimization Services company that provides all the services required to design and execute a conversion rate optimization strategy. WiderFunnel managed to increase the click thru rate on one of our highest-volume landing pages by 162% – and just on the first experiment! Their team of experts analyzed six major factors on our page: Value Proposition, Relevance, Clarity, Urgency, Anxiety, and Distraction. After determining where the landing page lacked in the various factors, Google Website Optimizer (a free testing platform offered by Google) was used to setup and flight the test.
Google Website Optimizer
Ah the miracle of multivariate testing – in a matter of one week, WiderFunnel was able to test 45 different combinations of landing pages, isolating 11 different elements of the landing page. Now that is efficient optimization. Additionally, Google Website Optimizer automatically determines the smallest sample size necessary to be confident in the results, so traffic is not wasted.
The landing page performance experiment broke down like this:
- Style: 5 variations dramatically altering the overall look and feel of the page. This included the background image, typeface, download button color, size, and call to action.
- Copy: 3 variations all aiming to increase clarity and the value proposition rather than simply state the product title
- Demo: 3 variations intended to increase the clarity of the product offering.
- Total = 45 combinations
Google Website Optimizer identifies the winning combination
- Style – the color palette was not originally relevant enough for the target audience, which is young and female. The original landing page was blue and orange – gender neutral. Although we knew the demographic was young and female, I suspect everyone was just a little nervous to go full on feminine. But that’s exactly where the test led us. One new style had a big pink download button, pink headline, and pretty swirlies in the background – very feminine. The second new design was also very girly in a less direct way. The design played to the current “Twilight” phenomenon. Style had significantly more weight than any other element that was tested. Three out of the four top-performing combinations had the same design.
- Headline – updated the headline to highlight the value proposition of the product. The new headline actually lists out the features of the products, and emphasizes the fact that it is free.
- Call to action – added the missing value proposition that the product is free to the download button (and significantly increased the size).
- Copy – the original copy actually performed the best. You can’t win ‘em all.
- Demo – this is referring to the graphic on the right side which is a flash demo previewing some of the product’s features. In order to improve the clarity of this demo, a mouse image was added over the buttons to emphasize their function.
Control

Winning Combination #1
Winning Combination #2
Validation
Internal testing on our Experimentation Platform confirmed the click thru gains seen in Google Optimizer. Combination 2 actually was the highest performer, once revenue was factored in with a 51% lift in earning per click.
Take-aways
Plan. Spend some serious time understanding your audience and then catering to their preferences. Don’t be afraid of big pink buttons. Speak to your audience as well. Put the value proposition directly in the headline – don’t make it hard to find. And keep the copy simple.
There are other tools out there, but we highly recommend using Google Website Optimizer to run multivariate tests for measuring landing page performance. The analysis and changes can certainly be done without an agency, but we also highly recommend working with WiderFunnel if your budget allows.
Kristen Groom, Senior Marketing Specialist, W3i
Kristen uses her experience in internet marketing, graphic design and testing to lead the charge in marketing management, optimizing all touchpoints with consumers.