The ConversionXL Process
Everything that can’t be explained as a process, is improvisation.
This week I was working very hard to finish the longest session so far in the program of ConversionXL Institute, of Growth Hacking and is the course of the ResearchXL model, taught by the founder Pepp Laja. Unfortunately, I could not finish the whole class as I reviewed several times the most complex videos. So far, the moral is that I must apply as soon as possible everything I’m learning in this program, to see the results of the ConversionXL model in my company.
Another motivation I have to apply this process in my company is that in my experience, you always learn a lot more from applying a theory. This does not mean that theory or processes are not necessary, quite the contrary. What it implies is that during the application of processes and depending on the company in which you work, you must adjust the processes, and eventually not all the processes can be applied, because of the value it generates for a specific situation. Again, it is extremely important to have processes, since what does not have processes, is improvisation.
Having said that, let’s jump to see what I learned during this week. As I said before, this week I was doing a Deep Dive in the four steps of the Framework to be able to optimize a web page. These were:
Four Steps Framework — Research Analysis
- Heuristic Analysis to assess user experience
- Digital Analysis to identify problem areas
- Technical Analysis to identify functional problems
- Qualitative Research to draw insights
In first place, the Heuristic Analysis to assess user experience. Go to every device that users use, to every screen of your web, and take the 4 feature tests. 4 features to consider when auditing your website experience:
- Clarity: Is everything as clear as it should be.
- Friction: What is holding me back in taking action.
- Anxiety: What causes me anxiety.
- Distraction: One goal, several goals, or a lot of distraction elements.
Secondly, Technical Analysis. Run a technical test to identify that every page of your website is running smoothly from a technical standpoint. See CR in every device/browser difference to identify possible problems. Cross devising Testing, Cross-browser Testing and Speed Testing
Thirdly, Digital Analytics. Digital Analytics works for identifying where the money is leaking.
- Identify the path of your website visitors
- Identify where your visitors leave your site (Delivery Points)
- Identify where the friction points are
- What kind of customers are the ones who go out
- Identify correlation behaviors
- Check if everything on your website can be tracked
Fourthly, Qualitative Research. Surveys and user testing
- On-site surveys (search on key pages of the website — with low exit rate)
- Follow-up surveys: Surveys of the people who actually pay for our product
In a second place, it is necessary to start with an analysis of the technical aspects, these eventually are seen as a “hanging fruit” because changes in technical aspects that are had in the page, can affect positively in the results that are had in the page in the short term. In this sense, it is important that before starting to work on these optimizations, we take into account the numbers that Analytics software can give us. For example, if my website has a higher percentage of visits from the desktop version, this should definitely be the device that I give priority over the tables or the mobile. However, this does not mean that I should leave the other options abandoned, it is more about giving priority to elements that have a greater impact on the business and in that sense have a hierarchy to attack.
Firstly, the questions we must ask for the technical analysis are: Does the site work with all major browsers? Does the site work with all devices? How is the user experience with each device?
In addition, we have to check the versions of the browsers and see if the conversion rates are similar between them. And check the tablet vs. desktop conversion, they should be similar. If the tablet conversion is lower, you may have a problem with UX in the smaller browsers.
What we need to do is to perform a thorough review of the site with all major browsers and each device category (desktop, tablet, mobile). Pay attention to the structure of the site, and definitely go through the process of filling out the box/form. The goal here is to put yourself in the customer’s shoes and see what they are experiencing. It’s also a great way to become familiar with the site and its structure.
- Walkthrough the site in question
- Take note of the structure and handling of the URL
- Check the URLs of the pages when you move through them
After this, we reviewed the way in which surveys should be done and the need to differentiate this so that they are effective, to the indicated public and above all that the process to which they are Qualitative survey approach: Must be done by a zero-sum analysis. Get all the qualitative answers and make clusters according to keywords or insights found. Then do a Quantitative analysis according to the qualitative data.
Bouncing betas: When research is made to a very small audience, not all the customers are going to fit in the research, so the answers would be 0.
Error in surveys: Mixing behavior questions with attitude questions. Others are questions that don’t communicate. Surveys too long (5 to 10 minutes long max.) An error of central tendency, fatigue increase, and people answer with neither agree or disagree, but you can’t go further in the analysis.
Selective Perception > Something customers agree with you, tent to automatic agree
Survey Customers via Email
Important to send out a purchase as soon as possible, from the point that your customers purchase your service/product. From 8 to 10 questions max, avoiding the fatigue of customers.
Usability Testing Vs A/B Testing
The difference between both options is that usability testing shows what issues are causing the user problem or friction to accomplish a goal, against a/b testing that shows the probability that A option is better than B, with a statistical significance. In the usability test, you only need some users. In an A/B test, according to the number of web visitors, you need a specific amount of them to validate a hypothesis.
The way to create a test on our websites is to make a Usability test to find the problems on the website. Later, create a hypothesis, make an A/B test and get a result in your hypothesis.
Is useful to identify:
- Where people click and where they don’t
- How far they scroll on any given page
For a Head Map analysis, it’s possible to use tools like “algorithm tools”, but it’s necessary to take into consideration that these instant head maps are created by machine predicting algorithms, so there’s no attribution to real end-users.
I will continue posting the next part of the class as soon as I complete it. I will be in touch soon.