The conversation about data governance has moved on from the why to the how stage. Bold suggestions on budgeting data governance activities are as routine as those for marketing campaigns and technology purchases. Gartner goes so far as to predict that by 2016, 20% of CIOs in regulated industries will lose their jobs for failing to implement information governance successfully -- and for that to happen, spending on that governance needs to increase five-fold.
I keynoted last month at the MDM & Data Governance Summit in New York, and if attendee questions and interest are any indication, data governance is further along than it was earlier this year, when I presented in San Francisco. Unfortunately, what's also clear is that this zest for data governance isn't yet matched by clarity of vision, or consensus on how exactly to go about doing it.
The most fundamental thing to keep in mind is that data governance -- like any form of governance -- is as much art as science. I've seen lots of definitions that cover policies, processes and people but miss out on another key dimension: politics. Like it or not, you're going to have to hone your political skills to line up initial and ongoing support for your data governance program.
And despite all of the talk, there's still confusion about what data governance is exactly. Is it about data quality? Metadata management? Taxonomies? Reference data? Master data management? Transactional data governance? Business analytics? All of those things?
[ Once you have a plan you will need people. Read Big Data Talent War: 7 Ways To Win. ]
Without a clear vision, your situation can be described by this proverb: If you don't know where you are going, any path will do. Or the related proverb: If you don't know where you are, a map is of no use.
Complicating matters is the fact that there's no single tool for data governance -- something that was evident from the vendor offerings and other presentations at the New York conference.
The problem isn't a shortage of products. There's the master data governance software from SAP. There's a suite of data quality, data analysis, RDM and MDM tools from Ataccama. There's data governance software from Collibra that, under the hood, looks more like a business semantics and metadata management tool. There's software from Titus that takes a different tack altogether, toward information security and asset management. There are confusing "data governance services" offered by the likes of data quality tool vendor Trillium Software. Ponder these products and you'll quickly realize why a "data governance solution" is more wishful thinking than something tangible.
Seeking another perspective, I reached out to Rob Karel, a former Forrester analyst and now head of product strategy for Informatica. (Disclosure: My company uses Informatica products.) Karel holds the view that data governance is a decision-making framework tied to a cross-functional organization that aims to optimize a company's return on its data assets -- so no single tool can provide the foundation.
For example, change management is core to data governance, and no single tool can help accomplish that. Yet, as Karel points out, many enabling technologies support end-to-end data governance and stewardship processes, including "data profiling tools, business glossaries, data modeling tools, data quality and MDM tools, data security tools, data integration tools, spreadsheets, email, collaboration platforms, network drives, wikis, business process management tools, BI and reporting tools, whiteboards, paper and pen, etc." Karel maintains that many of the common tasks performed by data stewards are being consolidated into a more unified stewardship user interface.
Vendors such as SAP and Ataccama have some nice governance capabilities, and Informatica's IDD (Informatica Data Director, the GUI sitting atop Informatica MDM) is improving with every iteration. Yet we're far away from having a reasonably complete solution for even the more structured components of data governance.
So where does that leave you?
In short, if you're looking to implement a data governance program, stop looking for a single, comprehensive solution. It doesn't exist. Instead, build up a portfolio of solutions, small and large, customizing your approach (and architecture) to your own organization.