Datalab, a misunderstood concept in its early days
When the Datalab appeared, in the mitan of the 2010s, companies jumped to the top of this new concept. For the sake of responsiveness, they immediately mobilised their Business Intelligence teams by asking them to “innovate”. But instead of innovative solutions, they mainly obtained decision-making reports and new dashboarding tools.
In fact, the Datalab is not just a label that would simply be stuck to an existing organisation. To work, this laboratory must bring together profiles from the world of data, but also UX/UI designers and business experts. It is only through this collaboration that a Datalab can really produce what is required of it, i.e. innovative solutions. When this philosophy is respected, results are achieved quickly. For example, it took just a few years for Axa’s Datalab to develop an anti-fraud app that is now used in several countries. The experiment was so conclusive that the insurer has now increased the initial staff of its laboratory five-fold.
The Datafab, armed arm of the Datalab
If, on its own, the Datalab makes it possible to innovate, it would be nothing without the Datafab, which makes it possible to industrialise new solutions. This complementary approach is based on a software architecture dedicated to data collection, storage and processing. As a general rule, this architecture is cloud-based, much faster to implement than an on-premise solution. This is all the more true when the application that we wish to industrialise is intended for ephemeral use.
In this respect, an example remains in memory. In June 2020, RATP adopted a new application to detect the wearing of masks in transport. Developed by the start-up Datakalab, this application could not have been created so quickly without the Datalab-Datafab duo. And even though the CNIL finally put a stop to the system, the experiment proved conclusive. Because on now knows that it is perfectly possible to innovate and launch an application in record time.
Data concepts: do not confuse speed with precipitation
Companies tend to rush into new data concepts because data is a strategic issue for them. But it would be wrong to confuse speed with haste. In order for a concept to bear fruit, it must first be studied and assimilated. Companies must make an effort to understand and implement correctly. In fact, when they started to understand the Datalab approach, the benefits were there. Previously negligible, the production deployment rate rose rapidly to around 20%. And it has been growing ever since.
Taking good habits now is the key to progress in data centricity. In the coming years, new concepts are likely to emerge. For example, in the wake of DataOps and FinOps, we could very well see the emergence of” could very well emerge. An approach that will make it possible to optimise server batteries in order to reduce the environmental impact of a data project. Only by focusing now on the concepts available to them will companies be able to integrate and apply the concepts of tomorrow more quickly.