tl8.io: the first learnings

I summarize what I've learned after launching tl8.io in an attempt to remember them better for the next time I launch something.

When I wanted to get traffic on tl8.io, I first tried to identify my target group. Who are my users? Some of them are developers who want to then introduce the product to another target group: project managers, testers, product owners, etc. TL8 also relies on ngx-translate, an Angular library. This is why I contacted the main contributor of this library to ask him to put a link of TL8's website on the Github repository of ngx-translate. Results: Out of the 130 unique visitors of the website, 48 came from Github. And more important, out of the 16 converted visitors (the ones who actually downloaded the app), 12 came from Github. I'm glad I managed to reach my target group.

Test thoroughly every new release

Yes, I deployed something broken. I didn't check all the iterations I deployed. That was a mistake: most of the users who tried the product actually used a broken version of TL8. Automation of such tests costs time and effort but they would free my mind of the fear of releasing new versions in the future.

Automate the deployment pipeline

This learning is tightly linked to the previous one. I didn't spend enough time to set up automatic processes. This means that I have to do manual work to actually deploy a new version. This is tedious and error prone. Unfortunately, some processes are not that simple to automate. Which is actually one more reason to invest time into it, when you think about it.

No feedback is the worst

I created a Typeform survey to get feedback from users. The link to this survey is displayed on the website and in the tool itself. It might be at the wrong places. Or users don't want to give feedback. I only had one response. There is definitely room for improvement. This might include having it tested by friends, family and colleagues. This is definitely on the roadmap.

More anonymous tracking

Overall, I don't know how users interact with the tool. I track one or two actions but it'd be very helpful to know a bit more about what's happening. The expected outcome would be to gather data to improve users' journeys.

The idea is good

Apart from technical issues, I only got positive reactions when pitching the idea of TL8 to someone else. Most of the time, after 30 seconds, the idea was not only understood but the value was clearly identified.

Let's see how the upcoming weeks turn out for TL8.


KM

Photo by Gabriel Ramos on Unsplash

Kevin Merckx

Kevin Merckx

Software Engineer, Real Full Stack Developer: from software architecture to development operations and programming.