Understanding user-behavior is a cornerstone of Conversion Rate Optimization and knowing how to glean invaluable information about users is a must. There is no shortage of options when it comes to data sources. But what data sources should you focus on, and how does each one of these sources fit, into the bigger picture of conversion rate optimization?
In this post I will be going through some of the data sources that I have previously used and give my take on, where each one fits into a CRO program and how they can be useful for different businesses and circumstances.
Heuristic Analysis
Prior to looking into any type of data, the first step a conversion rate optimizer should take is to conduct a heuristic analysis of the website, page, or app in question. But what’s the purpose of this type of analysis and what does it do?
According to Jakob Nielsen, Co-Founder of The Nielsen Norman Group and one of the most authoritative voices, within usability testing, “heuristic analysis is a usability engineering method for finding the usability problems in a user interface design so that they can be attended to as part of an iterative design process. Heuristic evaluation involves having a small set of evaluators examine the interface and judge its compliance with recognized usability principles (the “heuristics”).”
This entails looking at the entire journey, from ad click to landing page and checkout process, all from a user’s perspective. At this stage, you’re looking for congruence between business objectives and execution, i.e. does the ad copy and images that we’re using, match with the audience we’re targeting and how does it fit with our main objective for this campaign? And how about the landing page experience? Does it correspond with the expectations we might have created on the ad level? You’re basically looking at all the different parts of the journey, with a fine comb, looking for potential problems and conversion killers.
This process should give you an idea of some of the issues that might be affecting conversions, either directly or indirectly, and it will also provide you with some initial hypotheses that can later be tested. As mentioned, earlier, you should always have the end goal/main KPI in mind, when you’re doing this. Ask yourself, what am I trying to achieve with this landing page, etc., and are these elements helping or hurting my chances of reaching my goal?
One of the ways you can perform this type of analysis is by identifying the user goals and then come up with a bucket list of things, that have to happen in order for that goal to be achieved. The idea is that, once you’ve identified the necessary steps for success, it becomes easier to evaluate everything else. THis will enable you to highlight each element that might inhibit your main KPI, for that page, and deal with it later on. You could even go for an A/B test to check if it really is a problem.
There are many ways of doing a heuristic analysis, but I think this is a good starting point if you’re just getting into CRO.
Digital Analytics
Looking into quantitative data to find trends and patterns on a large scale is crucial – especially if the website receives lots of traffic. Google Analytics is great at providing insights into what people are doing on your site, how site content/features are performing, and how your sales funnel is performing.
It offers a variety of options in terms of data segmentation including:
- Traffic sources
- Device types
- Browsers
- Demographics
- Campaigns
And much more!
Google Analytics is incredibly helpful for CRO purposes because it lets you measure how users are reacting to each element on your site. The main benefits are getting to understand the “what”, “where” and “how much”, and having a much better foundation for conducting heuristic and qualitative analyses to uncover the “why’s”.
The sheer amount of data and reports available can be overwhelming which is why you should always approach Google Analytics with a specific problem in mind and questions in mind.
Understanding Reports
Using the QIA formula can help with providing clarity and a solid roadmap when approaching Google Analytics. The QIA framework – developed by Chris Mercer, Founder of Measurement Marketing – stands for Question, Information, and Action. The framework is based on the following rules: “You want to know what the question is that you want to get an answer to. Next, you need to know what information is needed in order to get those answers. Finally, you need to know what actions you’re going to take based on those answers” – Chris Mercer.
Once the potential “action” is identified, it’s time to go and find arguments – based on data – that either support or undermine that specific action.
Finding the Story
Whenever you open up Google Analytics, it should be with the goal of answering a question. Let’s say you notice that revenue has suddenly dropped but your transaction volume is unchanged, meaning that although you’re still getting lots of orders every day, you’re now earning less in total, and now want to investigate, why this is happening?
You pull up the e-commerce report and start looking at sales trends for the past 6 months, and it becomes clear that the overall number of transactions per. Week is fairly consistent. Next up, you decide to compare transactions per. Week with the avg. Order value, and quickly realize that, while transactions have stayed stable, for the most part, your avg. Order value has dropped more than usual, which in turn has led to a massive drop in total revenue. Now that you know this, you’ll have a much better chance of taking appropriate action.
The key point here, is to approach web analytics with a specific problem in mind, before you start digging and searching for answers.
Another important point to remember, is to not fixate on specific values when looking at metrics. Studying patterns instead of individual values, allows you to see the bigger picture about where your business is headed and the challenges facing you.
If you’re interested in learning more about digital analytics, then check out this post about how to discover business opportunities with data.
User-testing
User testing is about observing actual people while they interact with your website, app, or software. The goal is to get a first-hand look into the pain points and overall experiences people have with your site. This will help in making the shift from gut-based design – I think, feel, and believe this is how it should be done – to user-centric design – what do my users think?
How does it work?
You find a number of people that are willing to take part in a test, which consists of performing a set of predefined tasks on your website or app. It’s important that the tasks are clearly defined, in order to eliminate errors/confusion that might arise from poorly defined and unintelligible tasks. Check out this extensive guide on user testing
The following tasks could be used for a user-testing scenario on an e-commerce store:
1. Find product X from brand Y
2. Add product X to the shopping cart
3. Enable discount, using discount code xxxxx
4. Sign up for the newsletter
4. Complete checkout
It’s important to note, that the tests should be performed individually, meaning you should avoid – at all costs – having multiple testers performing tests in close proximity to one another. This is to avoid testers influencing each other’s judgments to ensure the quality of data collected, is as reliable as possible.
You should also make sure to use testers that resemble your own target audience, so if your target audience is typically women between 55-64, you shouldn’t waste your time recruiting 20-year-old men, because the way they experience your site, will probably be much different from your target audience.
Heatmaps
Heatmaps are used to graphically depict the user interaction with, for instance, a landing page or any webpage. They also facilitate CRO by highlighting the parts of a website that are being noticed the most and generating the most engagement.
Heatmaps can be incredibly useful in CRO because they can provide insights into whether users notice and percieve images, text and CTA’s. They can also help you discover could the following things:
- Which part of your webpage or landing page is getting the most attention from users?
- Are there areas, on your pages that are not being noticed at all?
Say you notice that people are spending a significant amount of time on a specific area of your page, which is great because this is where you’ve placed the main message and most important selling point for your product. But frustratingly enough, you also notice that no one is really willing to take action i.e. make a purchase, in spite of them having read/seen your main selling point. This leads you to the two following conclusions:
- Users are not able to understand my main selling point (MSP)
- Users are not finding my MSP convincing enough to complete an action
So now that you’ve established your hypotheses around this problem, your next step could be to start testing a different variation of your MSP and slowly start addressing the problem at hand, by testing alternatives.
There are a variety of heatmaps ranging from mouse movement heat maps, scroll maps and click maps.
Mouse movement heat maps, also known as eye-tracking heat maps are the least adopted heatmaps amongst CRO experts, mainly because of the underlying assumption that mouse movement equates to eye movement. A highly questionable assumption, both empirically and intuitively, since there isn’t any evidence to suggest that these are in fact correlated.
Conclusion
Doing conversion rate optimization, whether it’s formally working on a project for a client or just trying to optimize your own site, is all about embracing data. And not just quantitative data – even though you might feel inclined towards it. The importance of applying qualitative data to uncover the underlying reasons (the why’s) is equally important as identifying your what’s and where’s.
Start off doing a heuristic analysis, to get the full picture of the user/customer journey, then dive into your digital analytics to look for problem areas and performance gaps that point towards clear hypotheses. Once you’re done with that, it’s time to involve real live user testing to discover the underlying reasons for these problems and get a first-hand take on how real people experience your site or specific pages.
Feel free to share and comment, if you liked this post.
