Friday, May 30, 2014

Dresner's Point: Can You Have Self-Service BI and Governance Too?

I know what you’re probably thinking after reading the title of this blog post: the two are obviously a clash of interests, successful BI requires governance, there’s no middle-of-the road approach and one side will have to sacrifice its interests. You’re probably also thinking it poses a serious management challenge.
The dilemma requires some business intelligence wisdom, so I tossed the question out to the BI users, vendors and consultants in one of my Friday #BIWisdom tweetchats after a participant tweeted that he had observed a “quantum shift” to self-service in BI delivery this year.

I asked, “Self-service is important, but what about governance? Can you do both well?”

Their opinions:
- “Yes, you can, but it’s definitely a challenge. And self-service requires even stronger governance.”
- “You have to focus on value creation and shape governance to not get in the way of agility.”
- “Governance involves multiple parallel processes; and the processes need to support, not hinder, the business.”

Those processes and roles of governance include standards, policies and procedures for steering, organizing, implementing and executing BI initiatives. And it involves validating the data and ensuring data security – or as one of the tribe tweeted, “processes to stop people from doing the bad things they are tempted to do with data.”

One of the group commented that it’s important to have tools that “liberate data (for the right reasons); but there are too many tools that, at the same time, also expose the data to abuse.” Another person agreed, tweeting that departmental data discovery tools enable line-of-business user insight, but some centralized control needs to be maintained.

The group agreed that the major part of governance is security control. But they debated what should be controlled. Is it just a matter of who can see what, when, and having the proper security profiles to control that issue? Is it a matter of keeping data locked down until it’s requested and the user and data are vetted? Someone tweeted that this doesn’t seem to align with the intent of self-service BI functionality.
Another tweeted that It’s important to have democratized access to the data and not have it locked up in silos but freely available to the lines of business that need it. And someone responded that there’s nothing wrong with centrally controlling data to ensure proper usage. But another participant commented that there is currently no easy way to transform and load small amounts of data into self-service tools.

The tweets about the role of governance regarding security ignited an important question, especially when considered in the realm of self-service BI: Does governance also cover data quality? After all, bad data leads to bad decisions. The #BIWisdom tribe concluded that data quality will always be an issue in BI, and good MDM practices can mitigate the issue. Several tweeted opinions that it’s important to expose bad data. And one of the tribe pointed out that there is “no such thing as bad data; there is only bad information. The same data can be turned into useful information in a different use case.”

They concluded in agreement that the the central problem of governance is ownership of the data and the BI initiatives. And most companies don’t have a formal way of approaching this. Therein lies the crux of the matter. Where there is no ownership, there is no accountability.

Bottom line: From my years of studying successes and failures in business intelligence governance, I recommend that organizations first evaluate how their various stakeholders might use, and could benefit from, the information the data yields. Leadership need to consider risks and vulnerabilities along with advantages and then develop a comprehensive approach to governance – including self-service functionalities. It’s also important to keep in mind that, even in a self-service mode, the information/insights may be applicable and cross over to multiple areas of the organization.

As I’ve seen time and again, the best way to take this comprehensive approach is to establish a BI Competency Center (BICC). The top activities for a BICC, according to respondents in our annual Wisdom of Crowds Market Studies are analytical model development, database design/management and project management. I believe that user education also should be a primary objective for a BICC.
So, yes, you can do self-service and governance well. If the BICC (or other non-siloed governance mechanisms) is effective and relevant to the enterprise business, self-service BI functionalities won’t be a vulnerability.

Howard Dresner is president, founder and chief research officer at Dresner Advisory Services, LLC, an independent advisory firm. He is one of the foremost thought leaders in Business Intelligence and Performance Management, having coined the term “Business Intelligence” in 1989. He has published two books on the subject, The Performance Management Revolution — Business Results through Insight and Action, and Profiles in Performance — Business Intelligence Journeys and the Roadmap for Change. He hosts a weekly tweet chat (#BIWisdom) on Twitter each Friday. Prior to Dresner Advisory Services, Howard served as chief strategy officer at Hyperion Solutions and was a research fellow at Gartner, where he led its Business Intelligence research practice for 13 years.