A lire sur: http://it.toolbox.com/blogs/inside-erp/whats-the-difference-business-analytics-vs-business-intelligence-58672
Business Analytics (BA) is a close cousin of Business Intelligence (BI). Both are meant to help companies make better decisions by analyzing business data. The difference is in their methods, and in the general direction of their analysis.
In spite of the overlap and confusion, there are still some applications which most people would call BA, and some that are generally agreed to be BI. These clusters of cases are clearly distinct, even if the majority of instances overlap BA and BI.
Business Intelligence, the most common form, concentrates on data from the present and the immediate past, and drawing conclusions from that. Business Analytics makes more of an effort to predict the future using more complex tools relying heavily on anything from statistics to neural nets.
Both techniques have powerful advantages and equally striking drawbacks. Trying to predict the future, is both difficult and fraught with uncertainty, and this is why BA is associated with more complex, sophisticated tools. BI relies on existing data, either historical in the broad sense, or as close to current as the system can deliver. This is lacking when the environment is changing rapidly, but it is easier to analyze and, of course, can be used to make predictions as well.
Another important distinction is ease of use. BI is increasingly concentrating on the idea of “self service” where users can extract their own insights from the data with minimal or no involvement of the IT department. BA's techniques are more complex so the hands-on work is typically done by a specialist, often called a “data scientist”, who is trained in the specialized tools of the profession.
In all cases, BI and BA are bedeviled by extremely low rates of use in organizations. It is not uncommon for companies to invest in BI or BA and only have 10 to 20 percent of the potential users actually use the products. This is something vendors and user organizations are both struggling to address.
Getting the use rate up is a matter of better training for users, combined with formally adding BA/BI to workflow processes and creating a culture that encourages fact-based decision making. All of these are basically problems with the human culture rather than the technology, which makes them particularly intractable.
Of course, this is changing too as the vendors of BA tools work to make their products more user friendly by encapsulating the complex methodologies as much as possible behind friendly user interfaces that are heavy on graphics for data visualization.
It's important to keep in mind that at the end of the day, the goal of both kinds of systems is actionable intelligence, which can guide corporate decision making. Both BA and BI have proven effective in helping businesses make better decisions than conventional “seat of the pants” techniques.
Of course vendors have a stake in this debate over definitions, and each will loudly argue a definition that most benefits their products. Since these products have different capabilities and focus, it's not surprising that the vendors cling to different definitions.
The distinction isn't observed much in practice, and even in theory, there are many other definitions which draw the line in different places. Most people settle for the fact that both methods use analytical techniques to draw actionable insights out of business data and leave it at that.