You can increase sales or conversions by 400%, according to a recent MarketingSherpa finding, by testing and optimizing your landing pages. Use multivariate testing rather than A/B testing, and you can reap those increased conversions much more quickly.
Multivariate testing is a mouthful, but worthwhile to be aware of. Think of multivariate testing as running thousands of A/B tests all at once. It uses a mathematical algorithm to rank multiple elements within a single test.
For example, imagine you want to test these types of elements:
Multivariate testing allows you to test all of these elements (and more) at the same time. It works with test cells as small as a few hundred, so you don't need hundreds of thousands of users for accurate results. And it ranks each element alone and in relation to each other, something A/B testing cannot do.
A/B tests still have validity. If you've never dipped your toe into the testing stream, A/B testing allows you to start slowly, tweaking a single element at a time. At a global level -- say you're trying to decide whether to sell apples or oranges -- A/B testing is a simple, straightforward approach to determine the better of the two options.
But if you want to get more revenue from fewer visitors and you're not sure what elements have the most impact on conversions, multivariate testing works best. Here are three things to think about if you're considering running a multivariate test:
1. What pages should you test?
The Web pages that have the highest value are prime for testing. Look at the two or three hardest-selling pages. Are they pushing as hard as they can?
Other opportunities to apply multivariate testing include landing pages for email, ad, or search campaigns.
2. What elements on those pages make customers take action?
Every retailer has an idea of who the customer is, what they care about, and what the site needs to do to meet their needs.
Look at your assumptions of customers: Do they care more about price? Quality products? Being part of a community? Getting information before anyone else?
3. Test your assumptions
Anything that you think drives conversions should be thrown into the blender.
Should a big brand put a really average product in the hot real estate to push it harder? Or should you put a hotter product in the key real estate? Both have merit, but which works better?
Multivariate testing is the difference between fishing with a net and fishing with a line. The upshot? You get better results more quickly, an important factor when you consider that most things you test won't have a significant affect on conversions. And you have the added excitement of finding hidden elements that improve conversions (hidden, because they likely won't be the elements that you expected).
Add those elements together for an optimized page that increases conversions and lowers acquisition costs.
