Why test?

If not testing your visitor's experience then you're severely limiting the performance of your web site and the contribution it could make to the ROI of your website. 

The most successful marketers monitor, test and apply research data to refine their sales message and presentation to provide their visitors and potential purchasers with the information they need to complete the buying process.

Improve the productivity of your web content 

Testing is the key to enhance your visitor's experience, reduce associated sales costs helping to build and maximise your online income and revenue earning potential.

Productivity in this context is a measure of the efficiency of your web copy or sales offer in terms of the required response or goal against the actual behaviour and response of visitors.

Testing Methods

There are various types of testing environments:-

• A/B testing
• Multi-level testing
• Multi-variate testing

A/B testing

A/B testing is the most basic and simplest and probably the quickest to setup. A/B testing is popular as a web analysis tool as it is relatively easy to set up in software to serve different page versions to different visitors.

In the context of A/B testing of web pages, web page ads or other online sales offers, two versions of a layout or design are presented to visitors on a random but controlled basis. Their bahaviour in response to the page is measured based on the test criteria. These could be:-

• E-commerce sales figures
• Newsletter subscription sign-up rate
• Downloads or request for information
• Click through rate in a sales funnel process eg taking the next step in buying process

A/B testing is a traditional and well-used method in direct mail advertising where mail-order companies split their mailing lists sending out different versions of a mailshot to different recipients and measure the responses from each mailing. 

Method

There a many elements of a web page or ad that can influence the behaviour of visitors:-

• Page/ad copy
• Length of copy
• Page design/layout
• Price offers
• Headlines
• Font style, sizes, colour
• Site navigation

Generally, it is advisable to test various alternatives individually eg larger, smaller headline, longer, shorter copy, colour of headline, size of text etc. There are no hard and fast rules to the best combinations.

A successful headline could have its impact reduced by the use of a inappropriate colour. Testing the successful headline in various colours would provide more useful and reliable data.

Multi-level testing

Multi-level testing takes the concept of A/B split testing a stage further. It utilises the same process recognising that various alternatives of the same element could have very different responses.

Using the headline example outlined above multi-level testing would serve to different visitors:-

• the same headline in various sizes
• the same headline in different colours

Whereas A/B testing would test the headline itself a multi-level test would test various options of the headline.

Multi-variate testing

Multi-variate testing as its name suggests tests multiple variables of a web page simultaneously rather than individual elements. It can test multiple options of each variable and using sophisticated statistical analysis provides information on the most successful options and combinations.

In the example above a multi-variate test could test 6 variations of a headline in 6 sizes with 6 different colours. The main disadvantage is the exponential increase in the pages needed to test all possible combinations (46,656 in this example).

Any change to your web site can be tested to monitor the effect on visitor experience and site aims and objectives. If the change has a positive effect, keep it. If it doesn't, drop it and test another element.