Data quality and data governance just don’t get any love in the big data world. Compared with all of the technology innovations in Business Intelligence (real-time dashboards, mobile BI, advanced visualization), data warehousing (MPP architectures, in-memory computing, data virtualization), and advanced analytics (in-database analytics, Hadoop, NoSQL), poor old data quality just doesn’t seem exciting.  But that’s soon to change with the advent of “decision governance!”

Before launching into what I mean by decision governance, let’s do a quick review of what we mean by data governance.  I asked industry thought-leaders Martha Dember and April Reeve to share their definition of data governance:

Data governance (DG) is a formalized endeavor to manage the availability, usability, integrity, and security of the data employed in an enterprise. A sound data governance program includes a governing body or council, a defined set of procedures, and a plan to execute those procedures.

When I hear Martha and April talk about data governance, to me it translates into a management mindset with a set of policies and procedures to treat data as a corporate asset, instead of as a cost item to be minimized.  And that’s a good backdrop on what decision governance might mean.

What is Decision Governance?

Decision governance is about instilling a discipline for 1) measuring data-based decisions at the individual and organization levels, 2) managing and growing the organization’s decision-making assets (data, analytic models, decision-delivering capabilities) and 3) growing the decision-making effectiveness of the organization and culture.  Let’s examine each of these in a bit more detail.

Making Data-based Decisions

Getting business users to “trust the numbers” has traditionally been a challenge.  There always seems to be a bevy of excuses as to why users would rather use intuition instead of cold hard numbers to make decisions – the data is too hard to access, the data is too late, the data isn’t at the right level of detail, the tools to analyze the data are too hard to use, I just don’t trust the numbers, etc.

So what can we do to help enforce and monitor groups’ and individuals’ willingness to make data-based decisions? Let me count the ways:

• Measure data-based decision usage.  To borrow a line from John Smale, former CEO of Procter and Gamble, ‘you are what you measure.’  If we expect users to become more compliant in making data-based decisions, then we need to measure and monitor.
• Capture when users accept and act upon data-based decisions.  Do this so that we can ultimately measure the effectiveness of the data and the models that underpin those decisions.
• Capture the reason for the situations when a user rejects a data-based decision.  We need to provide a means for rejection, but we also want to put in place a process for capturing the reasons why they reject the decision (e.g., data not sufficient, data too late, data not correct).  Not only does this allow users to feel comfortable with rejecting the numbers, but it also allows the system to learn why the decision was rejected.  By the way, the effectiveness of the users’ decisions still need to be measured, whether they accepted or rejected the numbers.
• Create a Decision Governance Scorecard that measures:
  • The number of data-based decisions presented to the user
  • The number accepted
  • The number rejected and the reasons why rejected

Growing Decision-making Assets

Like any other corporate asset, there needs to be a Decision Assets Lifecycle process that manages, grows, and retires the different decision-making assets.  It is another corporate intelligence asset.  Each of these needs to be “tagged” and its effectiveness monitored in order to make intelligent decisions about their lifecycle.

Some of the decision-making assets that need to be managed include:

• Data including data transformations and data enrichment routines and algorithms
• Analytic models and the algorithms and data that power those analytic models
• Decision-delivering capabilities including presentation capabilities, user experience, and data visualization

Nurturing the Decision-making Culture

Probably the hardest thing to do is to change the culture of the organization with respect to not only accepting data-based decisions, but also addressing the organization’s unique cultural philosophy around knowing when and where to use the insights gleaned from the data.

For example, do employees know when and how to act upon insights gained from social media sites like Facebook or LinkedIn?  If you’re a Financial Advisor and you learn that one of your client’s daughters just got engaged from a post on Facebook, how should you act on that data?  If you contact the client directly and suggest some financial plans to prepare for the wedding, does that come across as being a “creeper” (a word that my daughter has educated me on)?  What is the culture of the company with respect to things like their customers’ (and employees’) privacy?  Certainly not easy questions to address, and ones that most organizations have likely not addressed, but these kinds of guidelines are an important part of the decision governance role.

So while data governance may lack the technology pizzazz that other information management capabilities like BI, analytics, and data warehousing are enjoying, data governance and its cousin decision governance actually are at the forefront of the big data movement. They are helping organizations build out their decision-making assets and helping them codify when, where, and how to drive data-based decision-making.

Subscribe to RSS