Quand un projet de Business Intelligence est lancé dans une entreprise, il est primordial de réfléchir aux données et à ce que l’on souhaite extraire de ces données. Cela peut paraître trivial, mais il est pourtant facile de se perdre en chemin face à la multitude d’outils BI permettant de traiter et mettre en forme les données.
Face aux puissances de traitement affichées, possibilités graphiques des tableaux de bord, et autres performances, il est indispensable de se poser d’abord la question des données utiles et de ce que l’on attend de leur traitement, avant de se lancer dans le choix d’un outil.
Voici un extrait d’un article de Craig Curran-Morton sur le site www.gantthead.com traitant de cette nécessité d’analyser les informations et de savoir ce que l’on veut en tirer :
In any business intelligence cycle, the analysis step is critical to providing value back to the organization. It is through this step that all of the collected data is compiled, sifted and distilled to provide the intelligence that is required by management to support the decision-making process. Whether you use manual or technology-based techniques to analyze the data, management is not looking for the data you gathered but the meaning of that data. The analysis of the data should be able to provide an answer to the question that management originally asked. A business intelligence function without analysis is like coffee beans without the grinder, filter and boiling water–they are still just coffee beans until something is done to turn them into coffee.
Part of this is based on the simple failure to define our terms when we fail to make clear distinction between information and intelligence. Information is only information, and provides us with little value on its own. Except for us to say: “Look how much information we have collected about our business!”
Intelligence, on the other hand, is organized, filtered, assessed and analyzed. Analysis takes the disparate pieces of information and makes sense of it. It is through the analysis process that we explore, understand and add the real value to the information by producing intelligence. If you look up “intelligence” in a dictionary, one of the definitions associated with the word is “the act of understanding, (of) comprehension.” Intelligence makes sense of the information, providing understanding and context and actively supporting management in their decision-making process.
Moreover, with analysis comes risk. The risk with this process is that we have to make assumptions and estimations about the information we are analyzing. We loathe doing this for two reasons. First, many of us define our worlds by the information that is in front of us. Sometimes, we find it difficult to makes leaps over the information gaps and feel compelled to fill in those gaps with the missing information. This is because we feel we need to have an entire picture sketched out in front of us before we can add the color. The reality of intelligence is that we will never know all of the information. We will never be able to collect every piece of data on any given topic and, if we did, it would take us far too long and the information would become outdated.
I would argue that business intelligence fails in some organizations because the analysis is either not done or not done to the level that is required to provide real value to the organization. Regardless of the techniques used to analyze the information and who does the actual analysis (person or machine), the value is in what you provide. If I hand my boss a stack of papers and she still has to go through the information to figure out what it means, I have provided her with little or no value. On the other hand, if I provide her with a two-page analysis of the contents of those stacks of paper, the value for me as an analyst increases–as does the value of the BI function I am advancing.
