Multivariate testing doubles sales
91.4% increase in sales from one multivariate test on USB-FlashDrive.com
How was this achieved?
Using Google Analytics we identified a SERIOUS problem on the ‘get a quote’ page – plenty of people were arriving at the page but most of them didn’t fill out the form. Instead they were lost forever (probably to one of the competitors)
If only we could improve the performance of this page we would have a serious impact on sales – visitors had already arrived at the site, seen the products and decided to click on the ‘get a quote’ button. If we could just convince more users to take that one final step and fill out the form the impact would be substantial and immediate.
We identified five key elements on the page that were most influential to the visitor and ran a multivariate test.
Multivariate testing is a powerful tool which allows you to test several elements on a page at once, and find out what works best on your website. It’s a bit like cooking; you can try out lots of different combinations of ingredients to see which recipe tastes the best.
Which ingredients did we test?
The five main elements (ingredients) we tested are highlighted in red boxes above. For each element we had two options:
- Telephone number – on/off
- Client logos – on/off
- Page title – 2 versions
- Form design – 2 versions
- Product information – on/off
All these combinations = 32 recipes!
When looking at conversions (sales to you and me) it’s incredibly important to focus your time on forms. They are critical in getting people to take that final step from visitor to customer, but are often put together with little thought right at the end of the web design project (sound familiar?)
So what was the winning combination?
The winning combination = 91.4% increase in quote requests!
After running the test for 47 days, the winning combination included a simplified version of the form which did not include the client logos.
This simple result gave our client a 91.4% increase in sales enquires, which in turn led to a massive increase in sales.
It’s impossible to correctly predict the best recipe
Even though we can make educated guesses, we can never be 100% sure what the best recipe will be. The only way to really know is to test! Testing our ideas lets us get the opinion of 1000’s of visitors and find out what really works best for them.
I must say I was surprised that removing the client logos resulted in an increase in conversions. The assumption is that having logos of recognisable clients would make people more likely to buy from you.
However in the case of USB-Flashdrive.com for their American traffic this simply wasn’t the case – removing the logos increased sales! Perhaps they found the logos distracting or annoying?
What’s even more surprising is that on the UK website (which has UK traffic) the logos had a positive effect and keeping them on the page actually increased the conversion rate!!
Expert opinion?
Multivariate testing truly gives you an ‘expert’ opinion of which design ideas work best. By ‘expert’ we mean the people who really count – your customers. Listening to your customers is crucial as they’re the ones who actually use your site and they’re the ones you’re trying to convince to buy.
