Baptism by Fire
Fresh from finishing his diploma thesis at Shanghai Jiao Tong University in 2006, Johannes co-created “the best Chinese/Japanese fashion store on the web”: everything was by the book, and the finished online shop was a marvel to behold. Riding high with expectations, Johannes faced the online store launch.
Unexpectedly, however, the launch became the day of reckoning as they realized their mistake: they had created the best product they could think of but had done practically nothing to get the word out about it. The website had zero traffic. What to do? They sold off all the stock almost overnight.
Their next idea was creating a company that could compete with Western Union. As is common with Johannes’ creations, this one was also by the book. Technically accurate with every information security concept you could imagine. However, on the day the money transfer business went live, they received a call from Germany’s financial authorities.
“They asked us whether we were stupid or something because we were building the perfect tool for money laundering,” Johannes explains, adding that they had to roll back everything and put it offline. “That was basically the time when I realized, ‘Oh man, maybe I should walk with somebody who can actually, who actually knows how stuff works,’” he quips.
These days of baptism by fire provided important lessons and led him to Rocket Internet, where he met Robert Maier, and together they co-founded LadenZeile.
“I think, as an entrepreneur, you’ll usually have more failures than successes. Usually, it takes about ten fails to have one success. That was no different with me.”
A Data-based Approach to Ecommerce
Fast forward to the present day and Johannes has been at the fore to propel home24 to glory and spread its wings across Europe. home24 is revolutionizing how the world buys furniture. The company is leading the change, moving home and living purchases from crowded stores and onto the world wide web. Since 2019, Johannes has been responsible for HR (which they call P&O — people and organization at home24).
“HR and tech is quite an interesting combination. And the way I look at it is that I try to contribute to building a smart company, which is a smart company consisting of smart people and smart machines.”
Among the areas he is involved in are learning and development, SQL for everyone, data analytics, and smart machines.
home24 is probably the only furniture company in the world that offers SQL courses to all its employees. That decision is informed by the company’s data-based approach to eCommerce. In this podcast, Tobias and Johannes converse in great detail about the centrality of data in any business that wants to stand out in today’s marketplace. A big data maverick, Johannes explains how his experience in data crunching experience, including at Ladenzeile, is helping home24 change how people shop online for furniture.
Tobias asks Johannes where a CTO should start when creating a data strategy. Johannes says the best place to begin is by asking yourself why you’re making it. Within the “why” question, the CTO should seek to understand what the data will be used for, how it’ll drive value, what decisions will depend on it, and what the main challenge is likely to be in data accuracy and consistency.
For most companies, data cleaning takes the most time. To this end, Johannes recommends that businesses seek to understand how the data will be used. This will inform you whether it is necessary to have it readily available every second or minute of every day or whether it’s still OK to have data that is one week old.
“That makes a huge difference for data operations. It also makes a huge difference in your architecture.”
“Having clean data is hard work. It requires… that you build sanity checks into your data processing pipelines very early on and that you follow the principle of fail early, fail hard. So the moment there is a sanity check (that’s) raising an alert, then you stop the entire pipeline because you cannot guarantee anymore that the data that you're processing is accurate. You would rather not show incorrect data and by that diminish the trust in your data.”