News & Updates


Growth hacking works: an anecdote from Klaro HQ

As we conduct many small experiments to make Klaro grow, one of them has worked just fine at absolutely zero cost (but 20 minutes of work). Frankly, I would not have bet on that one. Yet, that's precisely what growth hacking is about.

The Klaro team currently applies the growth hacking method with the hope of eventually turning Klaro into a profitable business. In short, the method calls for validating growth hypotheses via low cost experiments.

Since we started from scratch (we didn't build an audience before building Klaro as we first created it for internal purposes) we are currently focusing on the Awareness funnel step. One thing we want to do is to generate more visits to

Among various experiments in the current sprint (still in progress), I wanted to validate a hypothesis: could I make my (very few) twitter followers visit the website?

My experiment

I don't have many twitter followers, and I didn't use Twitter for years (until recently, to promote Klaro). All I know is that I acquired a few followers a while back when I was blogging about software engineering and ruby and working on

To test my hypothesis, I made two very small moves:

  1. I improved Bmg's README file on Github, and put a link to in the last section, to indicate that Bmg is used at Klaro.

  2. I tweeted a little bit about Bmg, and reacted to a few conversations about databases on Twitter.

What happened yesterday?

  1. Someone (ioverthoughtthis, the one I don't know) ended up discovering Bmg, either on Twitter or on Github itself. Thanks btw, if you're reading this, please contact me. I owe you a Klaro personal plan for free.

  2. That someone decided to put Bmg on hacker news, probably the top web site in the tech scene. The title was wisely chosen I must add: "BMG: A Production Ready Relational Algebra in Ruby". I'm not even sure I would have been smart enough to highlight "Production Ready"...

  3. And it made my day: BMG was on the first page for several hours, people upvoting it to 77 points (not too bad :-). A few people also shared the link on Twitter, including people with thousands of followers.

  4. That generated a lot of visits to github, 105 github stars for my little Bmg library, and 58 visits to Klaro's website...

  5. ... yet only 3 created Klaro accounts (so an acquisition rate of 5%), and only one page view per visit.

What did we learn?

Growth hacking is about validated learning (borrowed from Lean Startup). When small experiments work, it is important to reflect on what the team learned:

  1. Thanks to network effects one single tweet can generate many visits to our website. It "simply" requires getting the attention of someone with many followers, or who is willing to post on high-traffic websites. We just got lucky here, but that luck is important: we didn't cheat - it just happened. That says something about genuine interest (for Bmg and relational algebra, not for Klaro yet).

  2. There is a lot of interest in the tech scene for Relational Algebra. A lot of interest. We built Klaro with relational theory in mind, and with an innovative Relational Algebra library. That could be turned into some interest for Klaro itself. Experiments are worth conducting further down that path.

  3. People are not only interested in shiny new things: relational theory is 60 years old. Most probably because sound & simple theories are timeless. There is some beauty in them, and people love what's beautiful. That might very well yield further experiments @ Klaro.

  4. It is certainly time to revisit Klaro's homepage! That is, we certainly need to enhance the Acquisition funnel step now. A 5% acquisition rate is not that bad at our level, yet remember that it is from people who made the effort of reading an unrelated README to the very end. And people leaving the website immediately is known to be a problem!