Your web pages are your most valuable online real-estate. Adding a new feature and content could make sense on the element-by-element basis. For example, additional text can help your organic SEO ranking, a breadcrumb navigation can enhance user experience, product recommendations can help visitors make easier buying decision, etc.

At some point, your web pages will start to look like a Christmas tree after the holidays, which can adversely impact your conversion rates.

The reality is that you must do everything possible so that your customers can find the products they want efficiently and help them make easy purchase decisions.

However, at some point your efforts will become counter productive.

Is there a sane way of approaching this subject

After presiding to hundreds conversion rate optimization tests we can tell you with certainty that all best practices are dead. There is no general rule that can predict visitor reaction to a particular page component and even less on changes of multiple page elements at the same time.

So, do not even bother with all the articles that start with “7 best …” or “ top 10.” Such advices can work some and can completely fail for others.

What you need is a scientific approach that can give you specific answers to your specific web page and the exact set of page elements. This scientific approach is called multivariate testing. For those who have never heard of this term, multivariate testing is defined as the experimentation with multiple page elements at the same time.

Multivariate Testing Is A Scientific Approach

The reason is very simple. Each web page has its own visual signature that forces human eye to read its content in a different way. This is because our brains are wired to notice the most visible element on the page first and not to scan a web page top to bottom or left to right.

By adding or removing an element of a web page you will change its visual signature. This will then change the order in which elements are read, which can have an unpredictable impact on how this new version of the page is processed by a web visitor.

In the online business, your enemy is the very short visitor attention span. A typical reaction time is less than second or two. So, if you do not ‘hook’ your visitor quickly, he is gone and you have lost a conversion (or sale).

Be aware of A-B test flaws

One might argue that a simple A-B testing, where you will test the impact of one element at a time, will do the job. This is highly unlikely. Applying the results of the individual tests and then extending that logic to combinations of elements will only work if each element is completely independent from one another. This is rarely the case.

For example, in your A-B test you might determine that a red Add-To-Cart button is better than a blue one, and that green font color for the product price is better than a maroon color. However, after running a multivariate test of all the possible page combinations, you might find that the combination with highest conversion rate is the one that has maroon font color for price and a blue Add-To-Cart button.

Why such unexpected behavior? We already mentioned a technical answer: each combination has its own visual signature and conversion rate. Practical reasoning might be described like this: your product has a very completive price – the maroon color of your price is more visible against the blue Add-To-Cart button – as a result, the visitor is quickly ‘hooked’ to your price and the decision to buy is made more frequently.

The Usual Suspects

Here is a short list of troubled elements that should help you formalize your thinking and arrange multivariate testing:

We are sure that others can identify many other types of content or widgets that are being added to a typical web site. Our objective was to initialize your thinking so that you can apply similar testing criteria on any other web page element that your website might have.