Multivariate Testing for Website Conversion Optimisation. My POV.

Website Conversion Testing is a topic with many gurus, but few practitioners. And even the few tend to follow DM testing practices, without extending it to the possibilities.

There are reasons why this is the case - lack of support from internal Web IT Teams, limitations in the initial site structure, CMS limitations, tracking limitations, budgets, individual KPIs, need to obfuscate results, politics etc..But one of the key issues IMHO, is how the hypothesis for testing is being framed. So let's take a step back into the basics.

Why do we test?
The reason for testing is simple - to know what works. And use the learning to market efficiently. Period. Typical testing paths include Observe > Hypothesis > Test > Evaluate.

In situations where data is available, the hypothesis is mainly re-examining previous research & analysis than collecting new data. In the absence of data, it is usually a strong hunch with a potential big impact on business.

Common Elements of Website Conversion Testing
By it’s very nature, websites are perfect for conversion testing. Elements that can be tested include (treating each page as a Landing Page):

• Headlines – Content/Format (text/graphic)
• Offer
• Descriptive Copy – Bullet/Blocks. Lists of features & benefits and their ordering
• Presentation of Product/Service – Image/Screen Shots/Tours/Life Styles
• Call To Actions – Links/Buttons/Forms
• Elements that build confidence in the proposition – Users/Testimonials/Validation
• Contextual information – Links/Propositions
• Design Elements – Template/Layout

All of the above elements, if present on the page, can be tested across the dimensions of Relevance, Location, Quality, Prominence, and Proximity.

User Experience - an additional dimension to framing the testing hypothesis for website conversion optimization.
Besides the above elements that comprise a webpage, it is important to understand the layering of what composes the final user experience  from the abstract to the concrete, namely:

Layer 1 -    Site Objectives/User Needs

Layer 2 -    Content Requirements/Functional Specifications

Layer 3 -   Information Architecture/Interaction Design

Layer 4 -   Interface Design/Navigation Design

Finally, Layer 5 -     Visual Design

Today we have FREE & Paid tools to set up and run the website conversion tests, provided we are not hindered by limitations to testing listed above.  But it’s key we identify the layers impacting user experience being tested, along with the page elements, when framing the hypothesis to ensure the results can be used optimally.

To reiterate, the reason for testing is simple - to know what works. And use the learning to market efficiently. When website conversion hypothesis are framed to include Page & User Experience elements, the results should help you focus on the most important elements you should be tweaking, if any, to maximise the investment & learning. Happy testing!

Posted via web from Suresh Ramaswamy's posterous