Introduction
Conversion Rate Optimisation (CRO) is not about fixes or tricks. CRO is an iterative and continuous process. As long as you are working on optimising your digital consumer journey, you can improve results!CRO process can be broken down into four steps:
Step 1: Evaluate site performance
Step 2: Test hypotheses
Step 3: Analyse the results
Step 4: Iterate
Step 1: Evaluate site performance
Assuming your site has an Analytics tool in place (e.g. Google Analytics amongst the popular ones), qualitative tools such as Hotjar, Full Story, and the data is accurate, the first step is to analyse the current site data. We need to understand where the friction points are in the user journey and the conversion funnel. To do so, we need to understand users, their behaviour, and what they are looking for on the site.
1.1 Understand the friction points on the site:
First, it is crucial to identify the pages that represent the "weak links" on the website. Most likely, these pages are an open floodgate of lost conversion.The “Behavior Flow” on Google Analytics report is a good source of information for this step.Behaviour flow visualises the path a user follows from one page to the next or from one event to another. The reports help you identify the most engaging content on your website but, more importantly, identify problem areas and potential issues (drop-offs).
1.2 Identify the problem(s) with “weak” pages
Next, we need to identify the friction points on the problematic pages. What disturbs the user journey? To be able to answer that, qualitative research comes to help:
Heatmaps
Heatmaps provide a visual representation of the distribution of clicks on a specific page. By locating the clicks, we can understand how people interact with the pages and which part of the page they pay the most attention to. Red colour (hot) indicates the areas clicked the most. On the other hand, the blue (cold) areas receive few or no clicks.
Scroll maps
Scroll maps are a form of a heatmap that visualise how far visitors scroll on a page. Colour is used on the same principle to represent the most viewed sections of the page (red) vs. the least viewed (blue).
By analysing the heatmaps and scroll maps, we can identify issues in the user journey on our pages and likely find quick wins that will optimise the journey for our consumers.
Session recordings
Session recording is an excellent complementary instrument to have. This tool gives us granular detail on how the users engage with the site by recording their activity. The information can help us understand what goes wrong on the “weak” pages.
Once we analyse all the data available, we can then start creating hypotheses to test!
Step 2: Test hypotheses
Once we have identified friction points through qualitative and quantitative analysis, the best way to ensure our hypotheses are correct is to test them! To do so, we can use the A/B testing approach.A/B testing is an experiment carried out on the website or other digital channels (Ads, Mobile apps), validating hypotheses and potential improvements compared to a baseline. It allows us to see what is working for the audience based on statistical significance. There are other types of testing, such as Multivariate Testing (MVT), but this requires more traffic and is complex to run.The most significant benefits of A/B testing are:
Learn in-depth about the user’s behaviour in each test.
Remove the risk factor and subjectivity from decision making by adopting a culture of experimentation.
Focus on what works flawlessly for the consumers by learning from experimenting.
This phase requires a number of skills, including web development, in order to build the experiments and also run them.
Several tools exist on the market to perform these tests: AB Tasty, Optimizely and VWO (paid). Google Optimize also offers an accessible but minimal solution to test hypotheses.
Step 3: Analyse the results
Once the tests are completed and statistically significant, here are three potential outcomes:
The hypothesis is validated: In this case, the next step is to implement the changes that were tested. However, we should consider every aspect of the implementation, e.g. how much it will cost from a web development perspective. Thoroughly assess the technical constraints and the charge induced by the change.
The hypothesis is not validated: The test did not bring the expected improvements. In this case, we need to learn what went wrong and understand a failed test’s effect. If necessary, retest from the learnings.
Neutral result: the hypothesis does not lead to better performance. Similar to a test that is not validated, keep learning on the results and reiterate if necessary.
Step 4: Iterate
If you are satisfied with your test, move on to identify new avenues of optimisation to explore. This must be part of an ongoing process.CRO is a data-driven approach that leaves no room for intuition. The principle is to make decisions based on the data gathered. By integrating CRO into your digital strategy, you will improve ROI and lower your acquisition costs. However, CRO requires a lot of rigour and perseverance; it is the key to success!