Google Website Optimiser – Google Content Experiments
Although the route to increasing conversion rates on your website is both an art and a science, simply taking a random approach and trying out alternative headlines, pictures, and other items without any real strategy and hoping that that will work out in the end is going to leave you feeling somewhat frustrated. Take it from us; you need to discover some real data that will tell you what works and what doesn’t.
Fortunately there is a tool that you can use to test your conversions. The tool is called Website Optimizer and it is one of Google’s tools that if free for anyone to use.
N/B The tool has been discontinued and replaced with Google Content Experiments. A fuller guide to follow.
It might not be rocket science, but it is surprisingly effective and can provide some very useful results for very little effort. Essentially you set up a set of experiments that include all of the elements that you are using on your pages, and visitors to your website are served up the alternative versions randomly by Google.
One of the important benefits of Website Optimizer over alternative testing methods is that the tests are being carried out on visitors who are part of your target market. They have chosen to visit your website already, so what you need to optimise is the most efficient way to convert them.
Alternative types of test
With Website Optimizer you can carry out multivariate testing as well as A/B testing. Multivariate tests allow multiple pages to be tested simultaneously. This is ideal for testing different headlines, images, calls-to-action and so forth. The results are a little more difficult to interpret, but it is much faster than A/B testing of multiple elements.
A/B testing is the approach to adopt is you are simply testing one disparate layout or design against another.
Target conversion rates
Conversion rates depend on the kind of business you are in, though if your aim is to increase it then it is important to set yourself some targets. One approach might be to double your current conversion rate if it is higher than 5%, and set yourself a minimum target conversion rate of 10%. Naturally there are limits; if you are lucky enough to already be converting 20% of your visitors, aiming to double this to 20% might be overoptimistic.
There are many website elements that are known to have an influence over conversion rates, and these should be the primary items to test. They include:
- Your logo size and position
- Your headline size and position
- Colour schemes
- Headline text
- Body text
- Alternative typographies
- Calls to action
- Layout schemes
- Related products
- Labels and forms
But these are just examples; you know your web pages and the important elements. Make sure that you list all of those that are in any way relevant.
Which pages should you test first?
You also need to decide what the most relevant pages to test are; which will no doubt include your landing page, product pages, home page and so forth.
It is also important to understand how people use your website and where they exit. If most are exiting from your homepage then that it the page to concentrate first; if they are funnelling all the way through to your shopping cart before bailing out then that it the best place to begin your tests.
Setting up a test matrix
Now that you know the elements and the pages you need to develop a test matrix that puts it all together, for instance your test matrix will include alternative headlines, alternative layouts, alternative copy, alternative images and so forth.
Doing this will take some time, but with multivariate testing once you have assembled your test matrix everything will be tested simultaneously which will be far more efficient than running A/B tests over weeks and months.
Setting up the tests
In order to set up the tests you will need to have a working knowledge of html and the best source for detailed information on setting them up is Google; though you might consider getting some professional help at this stage.
Basically you will set up your test page and your conversion page. Next you will set up the content you will test. Then you need to select the proportion of your visitors you wish to test. Once it is all set up Website Optimiser will provide you with the code to insert in your web pages.
Some additional tips
- It is best to focus on moving your visitors through the conversion funnel step by step
- Site elements as well as individual page elements are also important
- Do you have different kinds of users? If so then you might serve them up different tests
- Are your competitors making changes? You can test those out too
- Big changes are more likely to yield results faster than small changes
- Test regularly; markets, people and fashion constantly change so make sure that you keep up.