Business Intelligence (BI, see both BI 1.0 and BI 2.0) does not really have a standard definition, it seems.  BI seems to provide a name to things businesses have always done.  Perhaps first done on "stone tablets", then on paper, then in spreadsheets, and now using BI systems.  Sometimes these systems are somewhat informal and done "on the back of an envelope" so to speak, other times these are highly formalized processes.

And just as there is no standard definition of BI, there is no standard BI system.  BI systems generally do not inter-operate.

How would it hurt business users if one BI application created by one software vendor could exchange information it contains with another BI application created by another vendor?  Seems like that would be a good thing.

That is where XBRL comes in.  XBRL is:

XBRL (eXtensible Business Reporting Language) is an open standard which supports information modeling and the expression of semantic meaning commonly required in business reporting.

Could XBRL be used as a standard way to model information contained in BI applications?  Is there a need for such a standard modeling approach?  I think the answer is yes to both questions.

Dan Bulos and Sarah Forsman of Symmetry wrote a white paper in July 1998 called "Getting Started with ADAPT" which highlights some of these issues.  ADAPT is a solution Symmetry created because BI applications seemed to have "an attitude" (as he puts it) or because each BI vendor implements their application differently.

BI is about more than "slicing and dicing" business information.  In another white paper called the "OLAP Council White Paper" written by Sarah Forsman in April 1996, it is pointed out that OLAP, the underlying engine on which most BI applications are built, and the multidimensional model, on which OLAP is built is more about flexible access to information.

What does that mean?  Well, what I think it means is that while those spreadsheets and Word documents were great tools for managing the volume of information we had to deal with years ago, they are really getting in the way with this massive amount of information we have today and that volume is increasing at a rate of 30% per year by some estimates.

Is there a better way than using spreadsheets and word processing documents and limiting ourselves to one view of most information because the information is locked inside those documents?  What if the information could be made more interactive?  The SEC coined the phrase "interactive data".  Perhaps the phrase "interactive information" might be more appropriate.  Either way, interactive is better than static and locked into one format.

What if the multidimensional model and its flexible access to information COULD be used for business reporting of all sorts, for example for financial reporting.  Gone are the limitations of the spreadsheet or word processing document.  Reduced (probably not gone) is re keying from one place to another and the errors caused by that re keying.  Because of this structured nature of this information, automated processes for validating the computations and other business rules can be created, replacing the manual task necessary today to be sure everything "ticks and ties".

What if we could make this work?  That is what XBRL and Business Intelligence is all about.

There are two videos which discuss this topic from two different perspectives:

How will all this work out?  Time will tell.

What does XBRL contribute to all this?  Could it be that XBRL is:

  • A global open standard format for defining a set of information (i.e. today each application does this in a proprietary way and you cannot effectively exchange information defined between applications)
  • A global open standard format for defining business rules for that set of information (i.e. this is proprietary today also, cannot be exchanged between applications)
  • The possibility of enforcing business rules during data input or at time of instance generation (i.e. if this is used in an automated process). This means that in the past, say, financial information did not have this global standard meta data nor was the information structured, it was unstructured; therefore in the past automated enforcement was not possible (i.e. there was no structure to the data, now there is)
  • A global open standard format for EXCHANGING those business rules for the set of information
  • A global open standard format for exchanging the information itself
  • Enables the ability to validate against business rules at time of data creation and at time of data consumption (i.e. because the business rules can be exchanged, this was not possible in the past because the business rules were embedded within applications and where application specific, therefore could not be exchanged)
  • A global open standard ETL (exchange, transform, load) export/import format
  • All this fits well into a relational model or a multidimensional model (i.e. applications we have today)
  • All this is extensible/extendable (i.e. the forms are not static, they can be dynamic, flexible)