Why Are We Here?

I recently came across an interesting online video entitled 'What Is Business Intelligence?' An interesting question and it does a very good job of introducing the history of BI.

The video was made by a guy in BI Solutions at Microsoft.

I suspect that it is designed to be a viral advert setting us all up for Microsoft's new 'BI 3.0' offering. Now that doesn't make what's in the video wrong per se. So take a look at the video and see if we can agree that the characteristics of 'BI 3.0' will be:

• built into the desktop (ubiquitous)
• work on structured and unstructured data
• combine private with public data sets
• be zero cost (at least built into the cost of your 'other' desktop software)

For me the next question is what does this mean for my area of BI - medium and large organisations? Here are my predictions:

• The mega vendors (IBM, Oracle, SAP) are going to be under much greater price pressure because Microsoft Office, DBMS products, the new more nimble (and much cheaper) vendors like Zap, Microstrategy, etc. and google analytics will offer much improved performance and tools than today. This reduces the 'value add' of the mega vendors.
• SAS will be less impacted because of its successful line of vertical applications and concentration on high-end analytics. Similar to the way Teradata concentrates on the big-end of data warehousing. SAS will still see some cannibalisation of sales at the low end and in government due to increased adoption of languages like R.
• Analytic appliances will become more commoditised as high performance universal data extract and transformation appliances - similar to how web servers are used today
• Microsoft will gain market share if (and only if) they simplify the accessibility of data from within the Office Suite.

So today, my BI platform is typically:

• High-end ETL tools: DataStage, Informatica, Oracle
• Mid-range database: DB2, Oracle, SQLServer
• Enterprise Analytics: SAP (BOBJ), IBM (Cognos), SAS

Tomorrow:

• Still high-end ETL tools because data (volume, complexity and number of sources) is expanding faster than the platform capabilities. The biggest change will be in extending the ETL function to include less and unstructured data from both internal and external (public) data from the internet
• Database will see me move more into SQLServer and open source such as MySQL because of the significant cost savings and the small amount of functionality that is surrendered
• Enterprise analytics will see the most significant cost change as the new vendors like Zap offer low cost (purchase and deployment) and ease of use.

Bottom-line impact? What costs me today $1 million plus in hardware and software licensing will be cut in half. That's a significant saving. For the small and medium sized enterprise the savings may be even more significant as there is greater opportunity to use open source databases and analytic/presentation software - something I would not yet risk in a major BI platform.

It will be interesting to review these predictions in 2 years time to see how good they are.

See you in 2 years time??