The 4thTh edition of the Paris Big Data Congress that has just ended has shown that the approach should not be seen as a mere marketing promise. The associated technologies are indeed a new lever for value creation, including in the field of Purchases.
Long skeptical of the contributions of the Big data buyers are starting to take a close interest in the subject. For example, in a recent study conducted by a large U.S. firm among 1,000 executives working in the logistics, purchasing and IT functions of very large companies, 97% of respondents recognized the benefit of this decision-making technology rising in their field. However, only 17% say they have implemented a big data solution.
The approach, which can be summed up in the definition of Gartner around the three "V's" (volume, speed, variety) by adding the concepts of "truth" and "value" of data, aims to facilitate the combined exploitation of information already held by the company, in more or less structured form, and those from multiple external sources, in various and varied formats. This, in an extremely fast way, by approaching treatments in real time.
To Purchases, whether it's analysing expenses more precisely, refining sourcing, strengthening performance management, optimizing supplier relationships, there are many areas of application.
"The available end-consumer data, typically used to meet marketing needs, is also useful for purchasing to build product profiles to be developed more in line with market expectations" Franck Le Tendre, Managing Director of SynerTrade.
And so to guide discussions with suppliers in the right direction, by renegotiating if necessary or by using specialists in other technologies. Of course, concrete examples of implementation are not yet legion. But, like most of his colleagues, Franck Le Tendre assures that "the purchasing departments are now in the middle of thinking to put in place a strategy to best manage this new paradigm."
Before making it more complicated, by multiplying the sources and cross-checking large masses of data, big data experts advise to start simply, with the information already in the company, knowing that less than 20% of them on average are structured. It is also recommended to start on a limited perimeter, by the most critical or simple shopping families, those that generate the most incidents, etc., and then to increase in power according to the results and means. Ambitious purchasing directorates can then go further, by collecting economic, geopolitical, meteorological, etc. data from public bases or on the Internet (websites, forums, social networks, etc.), for example to refine supplier risk management.
In terms of IT solutions, IBM, Oracle and SAP would be the main players in the market, followed by HP, Microsoft and a few others who are nibbling away. In addition, a number of publishers "encapsulate" the technologies of these hyper-specialists to offer big data layers dedicated to industries, business departments, etc.
E-buying publishers, especially those with a decision-making offer, have also understood the interest of positioning themselves quickly, rather by integrating external bricks than by developing their own layers. Especially since they must also manage another aspect, central to a big data approach: data management. Before launching analytical processing, all information must indeed be reprocessed (consolidated, sorted, cleaned, verified, formatted, etc.) to become usable and provide maximum value.