There has been one single thing that I could not get out of my mind ever since I started entrenching myself in research about Open Source Business Intelligence for a conference talk. I am talking about what kind of products should be considered “Business Intelligence” in a wider sense, and what shouldn’t. I learned from Wikipedia that Howard Dresner coined the term “BI” in 1989 to “describe a set of concepts and methods to improve business decision-making by using fact-based support systems”. This is a fairly wide definition, but arguably, your operating system should not be included in this set of concepts and methods. Jasper Reports* should not, either.
Open Source BI, and if you believe Forrester’s research, BI in general, is going to be quite a lucrative turf over the next few years. In times when niche growth markets are sparse, you will find a lot of 2nd tier players and some of the more obscure products getting stretched beyond belief to make them fit into the right market segment. Unfortunately, this is what I see is happening with Open Source Business Intelligence. This is a challenge for the rest of us, because it takes extra effort to weed out the wannabes from real solutions.
In my mind, a report generator does not constitute a Business Intelligence tool – it is merely a delivery platform for Business Intelligence. BI happens between the data and the report, and there you will find painfully few offerings in the Open Source arena. Mondrian comes to mind, and some of the tool aggregators that build on top of it, such as Pentaho. Maybe OpenI.
We should be a little stricter in how we define Business Intelligence. This is not about semantics, but about customer expectations that become more and more difficult to manage when commonly used terms are ill-defined.
* To do justice to Jaspersoft, they are about to launch a server-based OSBI stack in direct competition with Pentaho.
Totally agreed with you.
The true value of BI lies in transforming data into information, information into knowledge. Query and Reporting itself is not BI, only bundled together with OLAP multidimensional Analysis, Visualization, Scorecarding, and Prediction/Forecasting (using data mining) is BI.
The definition is not in tools or features but in support of the decision process. A simple spreadsheet can convey significant business intelligence that hours of complect multi-dimensional data mining may not. The test of BI is how well it responds, how agile it is in the service of human intelligence. Some BI tools are better than others, the best can “think with you”.
“BI happens between the data and the report”
I like this line – a succint and catchy but very apt description!