I was rummaging through my knowledgebase of BI stuff this morning and came across a curious news report on Gartner from late last year. Here's the bit that caught my eye:
So the jist of the claim is that search is not business intelligence (or ERP). I agree with the ERP bit but I think it is odd to split Search and BI. Sondergaard says "this new trend in business IT should not be confused with 'business intelligence' which refers to the collection, integration and analysis of business information to make decisions and forecasts". Sounds like the same reasons you use search in an organisation.
One prediction I am confident to make is that Search and BI will steadily become synonymous as practical distinctions disappear between structured and unstructured data. This is already happening today as better methodologies and algorithms emerge that have the sophistication to deeply understand unstructured data. You only have to experiment with Google to get how smart the googlesphere is getting.
Even my personal knowledgebase of BI stuff has become a vital part of my professional life. I use Devonthink Pro to manage a huge variety of BI stuff - currently running to over 250,000 objects from MS Office documents to podcasts, pdf's and 100 other formats. Searching it is a breeze and the ranking of results is very good. Good enough to make it one of the core tools I own and use.
Sure beats having a good memory (not that I have that choice!).
So here is my 2 cents on Search and BI:
- Both structured and unstructured data will be vital 'grist in the BI mill'
- Powerful algorithms will enable realtime mining of unstructured data into BI reports and dashboards
- Where such realtime processing is not feasible (because of slow response times or data integration bottlenecks) then batch processing of unstructured data to create structured data will happen
- Search of unstructured data within an organisation will be achieved by using algorithms developed and refined first by public use of search engines like Google.
- Mining public (or industry shared) structured and unstructured data will be as important as mining unstructured data within an organisation
- XBRL will be a major contributor to the structuring of publicly available corporate data in Australia - but as always we will be at least 5 years behind the USA in this.
Here is the whole Sondergaard article if you are interested. Download 200811_Sondergaard_on_Search
Comments