Quicker and better BI needed
The biggest dilemma facing organisations today is how to manage business intelligence (BI) quicker and better.
This is according to Barry Devlin, founder and principal of 9sight Consulting, during this morning's ITWeb BI Summit, at The Forum, in Bryanston.
“Data consistency and quality was the original business driver for data warehousing. In the past, there were multiple disparate and inconsistent data sources that were never designed to work together, and had incomplete and low-quality data. The goal was to achieve a single version of the truth.”
However, Devlin said there isn't a single version of the truth where functional business focus differs. “The changing world drives evolving measures; new ways of doing business drives change in data needs.
“Business wants a sufficiently correct answer soon enough to affect the outcome of decision-making. There needs to be a balanced structure between quicker and better BI.”
Devlin pointed out that advanced BI environments demand simpler, clear operational sources and an integrated design environment.
“If I can bring in the business people who understand the data, then I get something that's driven by their knowledge of the data sources. Shared metadata becomes easier to manage.”
He explained that organisations must start reorganising the data warehouse design and build teams to benefit from the knowledge that business has from acquired data.
“Increasingly, data is being distributed. We are going to have multiple databases and have to have the ability to virtualise access to those multiple stores. BI tools offer virtualised access to local data, data in marts, data warehouse data, and to non-relational data.”
Devlin added: “What I see coming next is that we will have an enterprise data warehouse but will become smaller than what we do today; there will be analytic stores, enterprise data warehouse and other stores; they will exist in parallel, with data virtualisation being the means to get access across them.”
According to Devlin, collaborative working and peer-based evaluation is critical to making BI more agile. He noted that self-service BI has a large element of hype in it because business users don't know enough about self-service BI and haven't used the tools to their full advantage. He said self-service BI usage must be built on top of IT-provisioned quality.
“Three tips for more effective BI delivery is to empower users to innovate; give them the power to play with data fast; have agile access and collaboration among peers.
“Focus IT on core data quality and performance, information and model reuse, and agile development. Bridge the business and IT gap using social networking tools, coring, ranking and peer evaluation.”