Insights / Site Optimisation & Design
Testing for Best Performance and Conversion
- 18th April
- John Lawson 10
John Lawson offers insights into the benefits of online testing and how to gain the best results to optimise site performance and conversions.
Always Be Testing (A.B.T), was a phrase that I learned very early in the game for my e-commerce business. In all my years of business, I never assume anything and test everything.
So how do you test? How do you get started? What elements should you test? How do you avoid the pitfalls? These are the questions that pop into most business owners and entrepreneurs minds when you think of A/B or multivariate testing. Lets dig in and take a look at what it is, how it works, and how to implement testing in your business.
What is page testing?
Testing is an experiment where you use a controlled environment and methods to try different versions of content on a web page. Ultimately you want visitors to take specific actions like sign up for your mailing list, liking a page or sharing it via social media or purchasing an item.
- A/B testing is a simple testing method of placing different variations of content and then determining which is most effective. One or more versions of a single page or page element are shown to visitors and the results are measured based on the defined test objectives. For example, testing two different pictures on the same page.
- Multivariate testing is a little more complicated as it is basically running several A/B tests concurrently on the same page across multiple elements. For example you would have two or more different pictures and two or more different product details, which would give you six different possible combinations to test on the page.
How do you determine test objectives?
Before you jump in and test you need to define your measurable objectives. For example
- Create Brand Awareness – visibility in your industry for your product or service.
- Increase Conversion – page views vs. actual purchases of the item.
- Reduce Customer Confusion – attempting to minimize customer questions or concerns about product or service.
- Improve Navigational Elemenst – using different versions of an ‘Add to Cart’ button.
But you do not have to limit this to functionality; you can also test the user interface and/or the look and feel of your site:
- Long vs. short copy;
- Call to actions;
- Buttons;
- Colors;
- Logos;
- Positioning on page;
- Images;
- Links;
- Page Layout; and
- One page vs. Multi-page checkouts.
Factors to Be Mindful Of When Testing
- Length of test period:
- Too short? – Be careful not to stop your test too early thinking your results have shown a clear winner, this can cause false positives.
- Too long? – That could be simply a time waster. But it also runs the risk of not producing accurate results due to fluxuation in traffic, seasons and demand. Plus the resources in use could be effective elsewhere
- Measuring the wrong stuff – this is a common bad practice for newbie testers. Sometime you can measure something downstream in the purchasing funnel that is affected more upstream at the top end of the funnel.
- Not taking proper actions after results – this happens when you get the results but do not make the necessary changes to capitalise on them. Also, use what is gained to proceed forward with more tests, its like shampoo ‘wash-rinse-repeat’ process.
