Posted on 13th April 2011 by Chris Lewis

Before I start, I’d like to make it absolutely clear that I think OLAP was a great technology and innovation, and when I first got introduced to it, I was blown away.

Although OLAP had been around a while, the first time I really started to become familiar with it was around the turn of the Century.

This makes me sound old but also shows how long OLAP has been around.

Coming from a relational background and spoon-fed on Oracle, Informix, DB2, Sybase and SQL Server I was firmly in the world of “Structured Query Language”.

I found great joy (and success) in learning how to do clever, unfathomable stuff joining as many tables as I could together in one go and exploring the limits of what was possible from one database supplier to another.

However, SQL was and is slow to retrieve data AND SQL is difficult and often a step to far for non-techies to understand and become proficient in.

When I saw OLAP and how easy it was to build a cube it came as a bit of a revelation. Why hadn’t I used this before? Why doesn’t everyone use it?

Well…

It works well because it’s working over pre-aggregated data.

“OK so there is MOLAP, ROLAP and HOLAP, all different variants of OLAP that either pre-aggregate, partially pre-aggregate or don’t aggregate at the expense of reduced performance… so if you want it to be fast it’s OLAP all the way.”

Building a cube takes time and over large data sets andfor really big cubes, can take hours. It’s this pre-aggregation that give you performance so that’s the trade-off.

OLAP only gives you half the story.

You can only see the data in an aggregated form and can’t drill through to the transactional data quickly as you have to go from the world of OLAP back to the world of SQL as soon as you go from a number to the raw data.

OLAP gives you flexibility through hierarchical dimensions and multiple measures but this is down to the programmer, the person who builds and maintains the cubes often hasn’t a clue what the business really needs and is trying to come up with structures that satisfy the many as opposed to YOU.

Move over OLAP and move over SQL.

There is a new way of doing things. A way that gives the business user access to the information they need that is quicker than OLAP, destroys the performance and flexibility of SQL and can be used by non-techies.

BI has been around for decades. There are so many products that do more or less the same thing in more or less identical ways.

Lipstick on a pig is a good description for modern day BI. At the end of day it’s still a pig.

A new layer of make up to provide a thin veneer over a fossilised architecture that is based on outdated ideas.

Search technology has been designed for speed. Designed for ease of use and designed for the non-techy (and it can be made to run real time…) and the 21st Century.

It’s about time someone thought about “Re-inventing Business Intelligence” by starting with a fresh way of thinking. By embracing the latest in search, social media, networking, web and phone.

When it comes to BI, move over OLAP and SQL, your time is up!