The problem is that suppliers make it all sound so easy when presenting to the higher ups. Our managers see a demonstration of some slick front end that shows everything they need to know about the business on one screen and they say "I want one of those". What everyone forgets is that the data to drive such a thing is not always available, or even is available from more than one place (then which one do you choose?). Point being, it's not just about fancy tools.

Couldn't agree more. Vendors make far too much noise generally - and thre's no consistency of message. Talk to one vendor and they'll tell you that a data-warehouse is the key to BI life, another will tell you its data integrity, another will tell you it's all about spreadsheet macros. We need one solution, a BI specialist that can actually come into an organisatio and provide an over-arching solution - nobody seems to be able to do that.

Granted, it doesn't particularly surprise me - since the data is coming from some many sources/ systems and trying to fulfill so many needs across the organisation as a whole. But unless you're prepared to go with a myriad of solution from a myriad of vendors BI generally is impossible. Has anybody actually thought about starting BI taskforce - a coalition of vendors all gunning for the same goal.

Actually, the more i think about the more impossible it gets. Lets have dashboards and Intrenets back - at least shifting everything online gives consistency of UI - we can worry about the rest gradually.

The entire field of business intelligence is a case of gobbledygook if you ask me. The idea that your business is some kind of starship that can be directly controlled from a central panel that tells you everything you need to know is just empty-headed and belies a trend in higher management to not to want to be concerned with those dastardly little details that can make or break a course of action. The thought that all of your business problems can be resolved by turning all those interrelated facts and figures into a couple of buttons and a check-box doesn't bear entertaining.

You can give me unrelated systems and managers who know their field inside and out over business intelligence any day of the week. Including Sundays.

There are good and solid methodologies that you can follow to arrive at a good and solid BI solution for your company.

ALL those methodologies start by looking at the requirements first, then model the data and then approach the construction of the warehouse.

The choise of reporting tools, dashboards etc are almost irrelevant if the homework is done there and then.

That in itself brings you to the root of the problem. Because requirement analyses, modelling and ETL can hardly be called sexy and it's a very hard sell, that aspect of BI is rarely sold. So the established BI vendors spend insane amounts of money trying to get you hooked on fancy reporting tools that might or might not fit your needs.

The only advice I can give anyone is that it's not up to any vendor to decide about what your needs are. You get less disapointments that way.

The subject of BI has turned into pure BS. New terms have replaced old ones but the underlying meaning is absolutely the same. A long time ago, in a galaxy not so far away senior management were always dependent on knowing as much about their business as possible; in those days the process was simply called "reporting". Other than the name BI is nothing new; the fact people are acting as if it's the greatest discovery of the century is funny.

The biggest challenge with BI is the attitudes of technical staff who do not understand business reporting and who are frustrated dealing with integration issues; the techies don't seem to realize that their paychecks are 100% based on the decisions management makes using BI. There is no reason for BI vendors to market to the tech crowd as they will have to deal with the system regardless of it's technical merits.

In some respects, the comments discussed in this article show that there is a lack of understanding regarding what the whole BI proposition is about.

Fancy front ends are all very nice, BI is very nice, but from the perspective of an IT manager, it can either be seen as a threat ("oh no, more work-- we can't do this - data is scattered etc etc") or as a positive ("hurrah - more money and now we can do things properly")

And it's the second opinion that should be being leveraged.

OK, so there is data laying in all sorts of different places. That's just pure bad architecture. How can you define "single point of truth" if you have multiple copies of data? How can you feel safe doing backups and restores if you have multiple, differing instances of the same data? How you you safely design future system or functionality enhancements?

Business"We need a system that does X with this piece of data"

IT group "Fair enough. Which of the 6 seperate sources of that data should we use?"

BI tools, especiallu enterprise level tools, are almost always linked to a decent data warehouse. And that's not "a big ol' DB", but a properly designed, dimensional based system.

OK, sometimes that isn't practical, but what is practical is using a "normal" relational database and putting data marts in front of it, which again are dimensional.

The net result, especially of the second option, is that have your data still intact, applications un-broken etc, but how have a single point of truth within your system as well. Which makes a huge number of IT tasks a lot easier. backups. Architecture. Design. Testing, BI, MI, auditing, compliance with pain in the backside issues like SOX etc.

If you have these, then BI front end tools are fairly trivial to implement.

The management have got their gadget and so are happy, but it's the IT team that have a swanky new data warehouse/set of data marts and a lot easier life in future.

Surely if the management want to give the IT groups lots of money to make their lives easier, it's a good thing, not a bad one?