Large organizations'•especially in highly regulated industries like financial services'•have an extreme need for data governance. The health and wellbeing of the business can be negatively influenced by people making decisions based on numbers that are not quality-controlled. At the same time, business moves very fast, making it nearly impossible for IT organizations to provide business users with all the data they need to get their jobs done. To make the decisions they need to make, business users go out and get data on their own and analyze it using spreadsheets and other productivity tools.
The question is: how do you create an environment where governance and agility co-exist? Where the data that must be controlled and protected is taken care of, and users still have the ability to obtain and work with their own data as needed? Where business users can ask'”and answer'”cross-functional questions?
A solution: a Business Discovery environment with analysis sandboxes
In the large financial services organizations he works with, Johan Averstedt often sees a scenario in which data wranglers harmonize and rationalize data for power analysts, collaborative users, and netizen users. (See the related post 'Self-Service BI: Power to ALL the People.')
The data wrangler's role is to determine if data is correct. These business users tend to be people who deal with risk on a daily basis. In banking, they may be in middle office roles like financial risk control or product control. In insurance, they may be actuaries. But they could be business or IT controllers in any part of the organization.
The data wranglers use QlikView to create an environment where they can conduct one-off analyses and monitor data over a period of time. We think of this as an analysis sandbox. (See related blog posts 'Analysis Sandboxes: Indispensable Tools for Insight Discovery' and 'Analysis Sandboxes the QlikView Way.')
The data wranglers invite a selected group of other business users to 'play' in the sandbox. They bring the data they want to work with into QlikView from spreadsheets, web feeds, etc., and the data wrangler spends time with the data to make sure it is correct.
Groups of people who perform this local analysis often come up with ideas that may be valuable for a larger group. These good ideas can then be channeled back to IT for use in bigger, more formal Business Discovery projects that are deployed to a larger community. Data wranglers may work with IT to make the data more broadly available to an appropriate set of users by creating a QlikView data file (QVD) layer from which QlikView apps (QVWs) can be built.
QlikView-based analysis sandboxes for data wranglers, and the power analysts and collaboration users they work with, deliver the needed combination of governance and agility.