Big Data is the future of everything, as you may have heard or read everywhere nowadays. Indeed, Big Data initiatives proved to be beneficial for many different companies, as we discussed here, and still hold potential for the future.
And with the business use of Big Data analytics growing so fast, small and medium enterprises are pressured to become more competitive in this field.
Last month Bpifrance organized a one-day conference to democratize big data and illustrate SMEs how to implement these solutions within their companies.
Paul-François Fournier, Bpifrance’s executive director of innovation, kicked off the meeting perfectly saying: “The theme for today is to dedramatize big data, since it can sound pretty scary.”
An intervention worth mentioning was the one from Franck Le Ouay, co-founder of Criteo and founder of Honestica, who said: “ It’s very important to start from the need and then develop the technology around it. In Criteo we had a simple idea, to create real-time advertisement that is personalized for the person watching. To do so, we had to collect a lot of data to understand what people liked. With this huge amount of domains, we collect, we can predict the probability of a sale, taking into account your navigation history. More than big Data, I would talk about machine learning, in fact, we need to, first of all, create the best advertisement, find the right price, which is the result of calculations taking account many parameters, and many other tasks that would not be possible without the machine. We could affirm that on average a person can reason with 4 dimensions. Working with 200+ parameters, it would be impossible to perform this task manually.”
Indeed, companies have always been collecting information and used it to make decisions, but something happens when these collections surpass a certain threshold, it allows you to see patterns and draw a conclusion, and that happens with the help of machine learning. It’s never about how much data you can collect but how much is enough to learn from. The actual benefit of big data is that we are able to use algorithms to detect patterns and take faster decisions, and fast means competitiveness. Therefore, machine learning doesn’t imply human replacement. Computers can find patterns but they don’t know how to interpret complex behaviors. Applicable insights come from human analysts (artificial intelligence is not there, yet?), but we still need machines to be able to work with incredible amount of domains, to polish these data.
Fournier continued: “Today we are at the level of maturity in this [big data] ecosystem that it can be used in every sector. There are examples of big data being used in agriculture or relocation. We need to get out of this “geek debate” and accept that it’s a subject that touches every sector. The impact big data on business is concrete. We conducted a study, enterprises using big data have growth factors that are 2x. On the margins, they experience growth from 20% to 40%. It’s almost like doubling the impact of the CICE, which is not something negligible. This has a huge impact also on the public finances.”
Companies are finally realizing the benefits of implementing data analysis technology and it’s important they do so from inception. The timing is perfect. Technology is getting cheap enough and its processes and architecture are optimized for enterprises. If you add in the mix Moore’s Law and the fact both big companies and some SMEs already gave proof of the benefits, you can see why we are facing a real big data movement.
The conference has seen some interesting testimonies from different entrepreneurs and big corporate professionals taking advantage of big data, as Yann Lechelle, COO of SNIPS, Bertrand Diard, founder of Talend, Xavier Perret, Director of Marketing at Orange, Sébastien Payen, founder and CTO of Fruition Sciences, Jean-Pierre Gauthier, managing director of Argus group.
You will be able to find all these panels on http://tv.bpifrance.fr/.