18 February 2007
Like a celebrity whose star is fading, business intelligence (BI) keeps trying to reinvent itself. And, like any human performer, it is unable to change its underlying nature.
Looking at the history of this class of software is like reading the Book of Genesis: MIS (management information systems) begat EIS (executive information systems) begat DSS (decision support systems) begat BI (business intelligence) begat BAM (business activity management) begat BPM (business performance management), ad infinitum.
Meanwhile, on the technical side, databanks begat DBMS (database management systems) begat SQL (structured query language) begat OLAP (online analytical processing) begat data warehouses begat data marts begat ETL (extract, transform and load) and so on, ad nauseam.
The latest example of this ceaseless titular inflation is to call business intelligence CI (content intelligence). Does this mean that business is now infra dig to talk about or is it simply the misapplying of an existing term?
It seems to be the latter. The idea of content intelligence is being promoted by document management companies such as EMC Documentum and Interwoven. They use it to mean classifying free-form textual data to make it easier to find and navigate.
In that setting, content intelligence is a reasonable, if typically overblown, term. I think it well wide of the mark, though, when applied to business intelligence.
Now, if the latest wave of BI tools were labelled “context intelligence” — and could meet that claim — I’d be much more interested, as would many organizations. At the moment, BI does not stray far from its origins in financial reporting and its reliance on structured data. It’s not bad at telling managers the ‘what’ of what is going on but hopeless at the ‘why’. Still less can it tell them the ‘how’ of fixing problems or repeating successes. The what, why and how, allied to relevant competitor and industry intelligence, would be information really worth seeing on one’s computer screen.
We are no longer accepting comments against this item. We suggest contacting the author directly.
20th February 2007: 'Eric Shayne Elliott' said:
"It’s not bad at telling managers the ‘what’ of what is going on but hopeless at the ‘why’. Still less can it tell them the ‘how’ of fixing problems or repeating successes. The what, why and how, allied to relevant competitor and industry intelligence, would be information really worth seeing on one’s computer screen."
I have to disagree. With the right tool, there is certainly the ability to dig into the data and get to a pretty solid "why". And if you have the right person using it, the "how" will often present itself as well. If you are suggesting that the program be able to spit that out without any analysis, well, I've seen plenty of executives that couldn't pull that off any better than the system would be able to...
Reply to Eric Shayne Elliott?
20th February 2007: 'Roger Whitehead' (Author) said:
> With the right tool, there is certainly the ability to dig into the data and get to a pretty solid "why".
I agree, up to a point. Human beings are brilliant at intuitive and inductive thinking and can pull sense out of a mass of data in a way machines can't rival yet, if ever. However, there's a limit to what structured data can tell you (assuming, of course, that it is first complete, accurate, up to date and consistent, internally and with other data).
What I'm arguing for is BI systems that surround (typically) numerical data with contextual material (numerical and textual) to make it easier for the less well-trained or intuitive to get at root causes. Such material would be drawn from broader and more varied sources than almost all BI installations presently provide. A system like this would also strengthen the hunches of the already adept data navigators.
At root, BI systems are almost all manifestations of a reductive and analytical view of business life. (It's no accident that many of them began in and for the accounts function.) That view is fine for certain kinds of thinking and decision but not all, nor possibly the majority.
Systems based on that approach do not present anything like enough of the right kind of information to provide all-round help to managers and other employees. If they did, organizations' networks would see a fraction of the telephone and email traffic they currently carry.
> And if you have the right person using it, the "how" will often present itself as well.
Given that the trend is to towards the so-called democratisation of BI, I would argue that the system should make the presence of "the right person" unnecessary. A system for all that demands a highly skilled expert to get decent results out of it is a failure.
> If you are suggesting that the program be able to spit that out without any analysis...
You know full well I'm not. 8-)
Regards,
Roger
Reply to Roger Whitehead?
