Big data is becoming a mainstream problem among the corporate world. In order to face this, companies have to start transitioning towards to a data-oriented model.

“Companies have to shift to a data-oriented model”


Interview with Johann Romefort, Seesmic co-founder, entrepreneur, semantic web architect and Big Data expert.

L'Atelier: Are Big Data and data analysis becoming “mainstream” problems for companies across industries?

JR: Yes, absolutely. Data analysis isn’t exactly new. It has been around for years but the recent development of cloud technology, captors and sensors brought the attention back to it. If you went to a conference about semantic web a few years ago, it was only hackers and nerds; Big Data was a popular topic among very specific communities. Now it is changing. I just attended the Strata conference on Big Data and I can definitely tell you that the corporate world has now entered the sector. Corporate and traditional companies are starting to understand that they have to shift to a data-driven model.

L'Atelier: What is a “data-driven model”, how is it going to change day-to-day life in a company?

JR: Being “Data-driven” means taking your decisions based on data. Traditionally, managers and CEOs would take their decisions based on what they wanted to achieve. Data-driven companies are focused on collecting data and analyzing it efficiently, and they take their decisions accordingly. Anytime they have an idea, they analyze data to validate this idea. All the major Silicon Valley companies are data-driven – Facebook, Twitter, Foursquare, Zynga… Data is the key to success and companies who make good use of it will be increasingly competitive.

L'Atelier: In order to do this, companies have to hire a data scientist?

JR: The data scientist will indeed become a critical position in the company. The data scientist is a cameleon, he has multiple roles and it’ll be a position more and more important in companies. The most important thing is that the data scientist understands the product perfectly. A lot of companies depend on commercial solutions. But the data scientists contextualize the data. Most companies right now use several commercial solutions, and try to esblish correlations between them. The data scientist, ideally, develops a single tailor-made internal solutions that unifies all the data. 

L'Atelier: Is it better to have data scientist than to purchase commercial solutions? Should small businesses have data scientists?

JR: Going from a commercial solution to an internal one is an extremely complex process and takes months. But ideally, a company should hire its own internal tailor-made solutions even if it’s costly at fist. An in-house solution can better determine what data is important for the company to focus on in order to bring changes to the product. Commercial solutions, on the contrary  tend to be very general, and depending on the business and the industry, companies will need more precise and custom solutions. Commercial solutions work for small businesses. But they still need to try and shift to a data-oriented mindset : start using some analytics tools, train their managers to data analytics – it’s absolutely crucial that managers understand this information in order to the, implement a transition to a data-oriented model. 

By Alice Gillet
English editorial manager