Introduction to Guerilla Data Governance: An interview with Mike Meier
Change Management In our continuing series of posts relating to data governance, we speak with data governance and data quality expert Mike Meier.
Mike has formulated an approach he terms "Guerilla Data Governance" and regularly presents at his blog and industry events his experiences of improving data governance without executive backing, budget and dedicated resources.
We caught up with Mike to find out more about his approach and some of the techniques he adopts.
Introduction to Guerilla Data Governance: An interview with Mike Meier
Data Quality Pro: Please can you tell our readers a little about your background.
Mike Meier: I have an undergraduate degree in Computer Science (and German) from Iowa State University and a Masters in Management from Cardinal Stritch University. I have worked in industries from Defense to Transportation including Manufacturing and Healthcare in roles from development to data administration and data resource management. I have publications as well as presentations on my resume.
Data Quality Pro: You are a strong advocate of “Guerrilla Data Governance” – can you explain what you mean by this term?
Mike Meier: By Guerrilla Governance, I mean the application of guerrilla principles to the creation of governance. Industry-wide (meaning the information industry) we have a tendency to view governance as a huge initiative that can’t be undertaken without the active support of the C-level and a substantial budget. This view causes us to wait rather than do. My own experience is that big results can be obtained without budget and without dedicated resources (though not without dedication).
Data Quality Pro: What are some of the main reasons you see for the failure of data governance in modern organisations?
Mike Meier: There are two main reasons why data governance efforts “fail”. First, the vision is unclear. The evidence of this is the confusion and uncertainty that abounds and the “wait to be told what to do” attitude that makes progress impossible.
Second, there are very real limitations on what can be done if we have no existing governance to tie into. Data Governance may be thought of as a process (a very complex one composed of many sub-processes) and if we recall our education in mapping process flows, we will note that the inputs to a process must be consistent and of known quality.
In many instances, we devote time, energy and money to defining the data governance process only to roll it out and learn that it takes human resources that are not available due to the number of decisions that must be made because of the complete lack of consistency of the inputs.
Data Quality Pro: Are there any techniques you have found particularly useful when trying to build a data governance community within an organisation?
Mike Meier: The techniques are those that we in technology-related fields have chosen to ignore. The core is Communication, which consists of listening, asking knowledge acquisition (rather than knowledge transmission) questions, and listening some more.
We have become focused on telling others what we need and why they should care—though we have never become adept at it. We need first to show others that we care about their problems and how, with their cooperation, we can solve them. Communities are built on shared concerns, shared pain. Every guerrilla leader has a feel for this.
Data Quality Pro: On your blog, you mentioned some informal data governance surveys that you undertook with members of varying sized organisations, can you summarise the findings for our readers and any conclusions that can be drawn?
Mike Meier: When I say “informal” I mean the kind of survey that one takes from the front of the room. Sometimes it’s a show of hands, and sometimes it’s a mental tally based on the kinds of questions being asked.
My audiences have included people from both very large and very small organizations. Because of the guerrilla governance subject matter, it may be that I’m not seeing people from organizations with “mature” data governance efforts.
I will say, however, that even in casual conversation at events from DAMA chapter meetings to the annual Data Governance Conference, there is an element of approval-seeking that feels like “are we doing it right?”
Data Quality Pro: You mention that in some of the “big name” organisations that are implementing data governance, staff are feeling a ”...considerable degree of anxiety…” or possess ”...a pretty fatalistic attitude…” about data governance – why do you think these people are feeling this way?
Mike Meier: I tend to think that the reason for the anxiety is that people feel that they’re in deep water and no longer in contact with the shore. The overall vision is lacking. Data governance seems manageable when it’s a matter of solving a handful of recurring problems, but once people leave the dock, the sea suddenly appears much larger.
The fatalistic attitude stems from a lack of ownership. In many organizations, initiatives like this are launched amidst much fanfare but with no feel for the commitment and dedication that must be present if it is to succeed. Politically-based commitments are like fireflies. There it is! Where did it go? Another reason for starting where the commitment level is highest—at the point of pain.
Data Quality Pro: Where do senior executives fit into “Guerrilla Data Governance”,the phrase implies that their involvement is not mandatory?
Mike Meier: I asked a CIO who was present how he felt when I said. “Just do it!” He confessed that it made him very nervous. I think that if executive management wants to be involved, it will only help and there’s no reason why one of these folks couldn’t lead a guerrilla movement.
The issue is that managers are accustomed to focusing on budgets and resource allocation and they operate in an environment in which political factors are a paramount consideration. This orientation leads directly to waiting for the stars and planets to align before starting something as far reaching as [data] governance. It’s not that they don’t want the results—they just don’t want to risk their credibility on something they really don’t understand.
I would love to have management be part of my community because they have access to information that others don’t and can help the community avoid situations that could be very costly in terms of credibility and commitment.
A guerrilla leader and movement can operate across a broad front in a very coordinated way so there is no reason to fear fragmentation and working at cross purposes—in fact, this is much less likely than under a more traditional approach.
Managers are forced to think first of their own budget and resources while the guerrilla leader is committed to a comprehensive vision and able to recruit new community members from other organizations as needed.
Data Quality Pro: Can any organisation start a guerrilla initiative or do they need a certain level of data maturity beforehand?
Mike Meier: Any organization can benefit from a guerrilla initiative. The concept was born in an organization that was a newborn in terms of data maturity.
What is required is a vision that has sufficient detail that the leader can define sub-initiatives and formulate a plan for attacking them, either in parallel or sequentially as opportunities present themselves.
At the start only one person needs this vision. As time goes by, other leaders will emerge as they begin to perceive the full scope of the vision.
Data Quality Pro: How is a typical guerrilla data governance team structured?
Mike Meier: There is no such thing as a guerrilla governance team.
Teams (I prefer communities) form themselves around specific problems or pain points. As individuals participate in these and solve problems (that management has either ignored or seemed incapable of handling), they begin to feel a sense of empowerment. Communities overlap and have no set source, format or size.
The leader(s) are always on the lookout for the “teachable moment” when they can take individuals or small groups aside and show them a little bigger piece of the vision.
Solving problems previously seen as insoluble or unimportant and relieving the associated pain creates a bit of hunger to do it again.
Data Quality Pro: In which type of organisational departments have you witnessed the most traction for data governance?
Mike Meier: Every department has recurrent problems characterized by poor quality inputs from some other department. These are incredibly frustrating for those who are involved.
When the only consistency in your day is your starting time, when you are being handed new problems at a rate that you can’t hope to handle, when you feel alone, powerless and stressed beyond your capacity—you are a starting point for a guerrilla initiative.
I think every department in every organization has those people. Their problems are real. They invariably involve data issues. Their solution is always a step toward [data] governance.
The only question is who is committed enough to step forward. The solution needs a focal point, a leader who can string successes together.
The leader who can see my pain as a data governance opportunity and use a win-win approach to create consistency and get rid of unpleasant surprises will be on everyone’s speed dial.
Mike has created the following presentation which also explains more about his "Guerilla" approach to data governance:
Is your organisation embarking on a data governance initiative? What challenges are you facing? Do you have any questions for Mike or our other data governance experts? Please use the comments section below to continue the debate.
Mike is launching a data management coaching practice as WhiteLake Data Management (http://www.m2dxtx.com). While a consultant may want to tell you what the answers are (sometimes irrespective of the questions), a coach helps you find YOUR answers and to recognize capabilities you may not know you have.
He maintains a blog relating to governance and coaching at http://bi-keep-it-simple.blogspot.com.
Useful Resources
See all posts in: Data Governance : Interview : Change Management
Michael Meier's Blog: http://bi-keep-it-simple.blogspot.com
Whitelake Data Management: www.m2dxtx.com
Applying Guerilla Governance Slideshare Presentation
"The Data Governance Imperative": Steve Sarsfield interview
DEBATE: How should data governance and data quality work together?
How to create a data quality competency center: Expert interview with John Schmidt
Creating a Data Quality Community: The Support Hub (Part 4 of 4)
Reader Comments (4)
Mike -
Here's a question: reading this after just looking at the Dalton Cervo interview(also here on DQPro) - how do you coordinate guerrilla efforts to avoid building up silos of information - it seems like, when you're doing this in the trenches, no matter how well you communicate within your circle of influence, you risk setting up one of lots of different data sources - and ultimately can end up with a project like Sun's where you're trying to weave together tons of disparate databases? Or are you advocating getting each different source to maximum quality and then trying to bump the effort up to C level to get it tied together?
M,
An excellent question--similar to what I heard from the CIO I referred to. The answer is that the kind of governance I propose isn't about creating information as much as it it's about creating good standard processes. When a group (community) begins to function at a higher level with fewer frustrations, it doesn't take long for the word to get out.
The communities I'm talking about might have a technology orientation, but probably not. In my success stories, the community members were users of technology and just sophisiticated enough to recognize that the problems that were causing all of their frustration were not related to the technology. They're not going to be aksing for meta data management or master data management or conformed dimensions, though the solution may begin to build a foundation for those initiatives.
Imagine a group of people getting consistent high-quality results in your organization. Could they hide their light under a basket for long? How much easier would it be to implement a meta data solution (etc) in an environment characterized by good, consistent process? Good process is more contagious than almost anything. "How do you do that?" Is the cry of the seeker. Nowhere in the interview nor in any of my presentations do I advocate or even discuss creation of information repositories. That can happen later when the pain is gone and the credibility is solid.
One last thought: If you want to be the leader of a guerrilla governance movement, you must be prepared to network into every part of your organization and digest a constant stream of status information--mostly expressions of pain in one form or another. You won't duplicate efforts but you will be able to duplicate solutions.
I like this interview very much. Funnily we labelled "guerilla applications" pejoratively as actually people in organisations that manage important data locally on their desktop and only share it ad-hoc, rather than sharing it. This gives them a reason to exist.
What is you opinion about the use of business semantics management (aka metadata management; see my blog about it) in the support of better data governance ? Communities indeed have shared ideas, pains, but also business concepts.
Pieter
Pieter, I'm sorry I missed this question until now. I think management of semantics (a kind of meta-data) is necessary. The only question is how best to accomplish it. Certainly whenever you present some information in the form of a table or chart, you must be prepared to answer questions about the origin and meaning of the data. Moreover, you will need to identify not only the intended meaning but also any problems or issues that might be acting to change, confuse or dilute the meaning.
Some of this information can be attached to the data items as, for example, Business Objects does during the creation of a reporting "universe". Some becomes part of the table or chart as titles, labels, notes, etc. Some is passed on orally because it may be deemed too sensitive to be written.
My only recommendation is: Don't try to do it all. Make use of the 80-20 rule which states that 80 % of the value is derived from 20% of the effort. Find the 20% that is going to give the bulk of the value and concentrate your efforts there. Anything else that you do will be mandated by political rather than technical or scientific considerations. It would be heaven on earth if we could agree from industry to industry and company to company on a minimal subset of metadata to capture. Unfortunately, this is affected by linguisitics as well as cultural forces and the span of those forces may be quite small.