Blog @romenrg

Thoughts and experiences; about Software, Internet and Entrepreneurship.

Linking UX and Conversions by Means of Analytics and A/B Tests

Thanks to my experience as Entrepreneur and Product Manager, I realized long ago about something that is totally counterintuitive for many other software engineers. This realization is that, when developing a software product, our main focus should always be sales. Always. Sales. I know that probably too many people have said that already, but it is so true that it is worth saying it once more, specially for my software mates: “when developing a software product our main focus should always be sales”.

However, that statement is too generic. Most people already know that selling is important… But how can we sell more? Well, experience tells me that in order to increase sales, the first thing you will probably need are many more visits (and getting those visits is way more complicated than it seems). However, that is only the beginning. Once you get enough visits and you validate most of your product main hypotheses, the next step is converting most of those visits into customers. Here is where both User Experience and KPIs (Key Performance Indicators) become probably the most important areas for product management.

User Experience and Conversion Rates

Every entrepreneur wants to convert visits into paying customers. There will be KPIs that you will probably always be checking, expecting them to improve. This KPIs could be “Signup rate” (from visits) or “Free to Premium rate” (the percentage of users becoming premium after being free users (in case you have some sort of free trial). Conversion Rates are very important business metrics, but improving those KPIs has to do with a few more things than just business. Particularly, the User or Customer Experience (UX / CX) is one of the most important factors in this task.

I always define User Experience as a discipline in which Design, Usability, Marketing, Psychology and Engineering merge. Are you wondering why? Did you think that UX was just “beautiful design”? Well… I have learnt that your experience while using a product is affected of course by the aesthetics (art and design), but also by the easier it is to use it (usability), the story it tells you (marketing), the way it engages you (psychology) and last but not least, “the engines”, that is, the technology that makes the product actually work and perform properly (engineering).

Once you agree with this definition of User Experience, linking it with conversion rates is rather easy… The better the UX in our Landing page is, the better the signup rate will be. In the same sense, the better the UX in our free version of the product is, the better the “Free to Premium” conversion rate will be.

But… Is it easy to constantly improve UX in our software products?

To be perfectly honest, no. Improving UX constantly is really complicated. And actually one of the most difficult parts is subjectivity, because this are the sort of tasks in which everyone in the team wants to have a say. One day everyone wants to change the text of the main headline to their own “best” option, wasting a lot of hours in useless discussions where nobody agrees. Another day some engineer named Joe says “it is imposible” to a new important feature coming from business that may be an important source of income because it will imply a few changes in “his” domain objects and he thinks it is not worth it. The same day, the back-end engineer that is temporarily working in the UI development, named Jane, decides to use red for the main CTA because is her favourite color, although most people associate it with “danger” and thus, will reduce the click rate… Do these situations sound familiar to you?

I have seen this behaviors several times in several different software companies and I have realized that they are caused mainly because two reasons: 1) most companies do not value User Experience as much as they should and 2) most companies are lacking data that would slap the corresponding people in the face and lead to the right changes to really improve UX.

“If we have data lets look at data, if what we have are opinions, lets go with mine” - Jim Barksdale

In my experience as CEO of Langproving and Product Manager of Vocabulary Notebook I have followed this quote as much as possible to avoid those discussions. To base our decisions in data, we have used both Web Analytics (tracking all the actions / events of our users) and A/B Tests (to compare objectively two versions of a portion of our product). To gather data and create our experiments we have used Google Analytics extensively. It has been a fantastic experience that has helped us, among other things, to significantly improve our conversion rates, as you can see here. Besides, those useless discussions mentioned above do not happen anymore in our company. Now we look at data. We listen to what our users say and we choose what they like the most.

Regarding the link with the results of some of our experiments, it is worth mentioning that recently I gave a course at Fundación General Universidad de La Laguna exactly about this topic (UX + Web Analytics). In this course, we had a first part focused on User Experience and a second part focused on Data (including our own experiments as examples). All the slides of the course are available (in spanish) in the following links. However, keep in mind that they are just slides that I created as visual support for my lectures during the course. They are not a book. This means that since they were intended to be seen along with my comments and explanations, they do not have much text and thus, sometimes can be a bit difficult to follow on your own.

Slides of my course about UX and Web Analytics