What Is CRO?
Conversion Rate Optimisation (or CRO as we will call it for convenience) is the art and science of improving how efficiently your website achieves it's goals – the goals being anything from driving sales through to lead generation. It covers a wide range of activities from off site (SERP copy, advertising messages, email copy) through landing pages and purchase funnels, and on into post-sale conversion areas such as gaining micro conversions on confirmation pages or increasing the chance of re-engagement.
To expand further on the art and science statement, let's look at the two components individually for now:
Optimisation is all about measuring results and continually making improvements. By tracking and analysing the results of your tests you can determine a definite success rate (and ultimately the impact on your bottom line) but also gain insight into what worked and what didn't, helping you to create the next iterations of your testing process and understand the dynamics of the relationship between your potential customers and your site.
There are many different methods of testing which feed into CRO, but at the heart of it is a/b and multivariate testing (using tools such as Google Website Optimizer) which pit your current site and revised versions against each other to confidently determine which is the most effective design.
The success of any process is limited by the creativity and quality that goes into it. Only by creating great, well thought out designs and copy can you expect to achieve an optimal conversion rate. It is impossible to test every variation possible to determine and absolute result – you need to have done the groundwork and figured out what should (and as importantly, shouldn't) be tested before you can even think about diving in to do any sort of testing.
An Introduction to Testing
The numerical outcomes of CRO are how we measure success and failure – to do so requires a solid understanding of the methods involved. To start this series which will cover all aspects of CRO, lets introduce the basic testing elements involved:
Testing two versions of an element or two different pages against one another to determine which version is most efficient at encouraging a task completion (sign-up, purchase etc).
More complex systems are tested using multivariate testing. Using a tool such as GWO, the users are served combinations which vary any element chosen to be included in the test, from copy and images, through to placement and formatting of the content. By statistically looking at the performance of all the combinations, the test will result in a winner once enough data is collected. Due to the amount of combinations included in the test (in contrast to the two in a/b testing) a longer amount of time is need to find the statistical significance which determines a winner.
There are no absolutes when it comes to conversion rate optimisations, human behaviour is a complex beast and each unique visitor's behaviour can not be predicted. However, by measuring a large enough set of visitors and actions, we can confidently predict that the pattern of behaviour we can expect each variation to cause. This is reflected in the way testing tools present results, in terms of the level of statistical significance – the probability that the outcomes are not just caused by a statistical accident, and are there for meaningful.
Now we have an understanding of the outcomes and goals we will get at the end of a CRO process. In the next instalment we'll look how the various components involved in CRO work together to produce the final results, and how you can make them work as a cohesive strategy to ensure a successful CRO campaign.