Framework for Conversion Rate Optimization (CRO)

By no means is conversion rate optimization (CRO) a new concept. Yet, many organizations, big and small struggle to follow a disciplined approach to testing, finding it hard to understand the benefits of the sometimes-expensive tools commissioned to achieve this objective.

CRO encompasses the optimization of design and content that lead to more conversions, whether macro or micro. Following steps are essential to create a sustainable optimization strategy.

1.   Identify a Starting Point - It has a few components to it.

a.   Business Requirements - What does the business require? What are the business goals? Optimization goals will always be aligned with the business goals. E.g. my business goal is to increase online sales, the KPI of which will be conversions/sales of online product.

b.   Data Gathering – One can utilize various data points, like online data, customer service calls and/or customer surveys. Online data will also help you decide on how you will deem an optimization test success (or not). What is my baseline conversion? What are the key pages that lead the customer to convert online? What type of calls do you continue to get despite a self-service option on your site/app? Going by the human-centered approach to solving a problem, ask your customers via a short survey. As Avinash Kaushik stated in one of his posts, the survey could be as brief as three questions: Why did you visit our website today? Were you able to perform the intended task? If not, why?

c.   Test Ideas – Using the data gathered, you can pick up pages/steps that are making the process longer/difficult/ unclear for your customers. Pick up pages that have high enough traffic to provide you with sufficient data points for statistical significance and confidence level. Adobe and Optimizely provide sample size calculators to know beforehand how much traffic you need to arrive at the chosen lift in conversion.

2.   Form a Hypothesis – Now that you have some test ideas, start with a hypothesis. If you make the registration process simple by asking for email ID only, rather than detailed preferences or if you state the benefits of subscription upfront, you are likely to get higher subscriptions. You may also want to include a specific customer segment to test. The three questions that must be answered at this stage are: What to test, who to test for and where to test.

3.   Ensure the house is in order – Before you start executing tests, ensure that the tracking exists and is accurate for you to benchmark current conversion rate and report on tests’ performance.

4.   Execute Test Ideas – Play with the layout: what elements contribute or prohibit conversions, whether placing a CTA higher on the page leads to more conversions. Test variations in the number of steps of the process; test different layout and content on the pages that assist conversions.

5.   Analyse Test Results – After the statistical significance has been reached with intended confidence level, it is time to analyse the performance. The best practice is to let the test run for two weeks, with the data being consistent for at least seven of those days. While interpreting test results, look deeper to understand if the behaviour is different for a set of users.

Other Salient Points:

1.   Optimization is a Process. It should not be treated as a one-off project. Rather, organizations should work towards institutionalizing it.

2.   Create a Learning Repository of the lessons learnt from the tests. Even when the test is not a success, it still provides important insights. It will have two-fold benefits:

a.    Not only can one department learn from the past tests, different departments can also learn from each other.

b.   Looking at the test results of a combination of tests may give an added perspective compared to a test result in silo.

How have you institutionalized optimization in your organization? What are the challenges you faced/are facing?

Please inbox me if you feel this is an area where your team can use help.