Data governance suffers a bit from the "chicken or the egg"syndrome. People at your company aren't going to understand what data governance is and what it can do for them until they actually see the results. However, getting the initiative funded and launched will only happen if you can convince your company of the tangible benefits of data governance. That can be difficult when there's no actual program in place.
Selling people on the benefits of data governance (a program that's going to cost money) calls for some persuasive salesmanship and well-crafted internal marketing.
If your company has suffered from a horror story that could have been averted with data governance, convincing senior management of the need for data governance won't be too difficult. Tell the horror story a few times, embellish a bit where needed, and funding should be easily agreed upon.
An obstacle to governance can arise if your company has already tried to establish a program and failed. Perhaps the business case wasn't strong enough, management wasn't convinced, and everyone was sent back to the drawing board for another try.
Regardless, making the pitch for data governance has to come from an unmet need to confront the results of poor business practices. One way to approach it is to research the avoidance cost: "How much is it costing us to not have data governance in place today?"
This requires spending some time away from your desk. Start by visiting business owners and gathering their data-related frustrations and pain points. Ask them for anecdotal evidence of returned mailings, orders shipped to wrong places, unhappy customers, fees and charges from third-party logistics companies, overstocked inventory due to duplicate part numbers, regulatory compliance issues and fines, etc. Once you get business owners talking about this, either individually or in small groups, you'll likely uncover a torrent of unmet needs.
Be prepared for the question "What are you going to do about it?" Explain that you're gathering information to build a business case for a data governance function within the enterprise, and that every story helps. Wherever you can, put a dollar amount next to the pain point. Even if it's an estimate, this evidence will support your hard-dollar answer to the question "How much is it costing us to not have data governance today?"
Based on the information you gather, use a simple survey tool like SurveyMonkey or Zoomerang to extend your reach. Structure your questions thoughtfully to get at people's overall business goals and how data governance can help support them. Make sure there are enough free-form answers in the survey so people can voice their ideas or frustrations with ungoverned, siloed information throughout your enterprise.
Ultimately, data governance will not be established and primed to flourish without the consent of the business and process owners with real power in the company. You need to recruit them from the beginning to back the establishment and fund the data governance function. If they don't see the benefit, or if they're actively against it for some reason, a data governance program will never get off the ground. That makes it essential to define a clear mandate of what data governance will do within your organization, making sure that the vision contains enough benefits so even the skeptics agree it's worth doing.
You'll have to live up to that promise. The initial data governance framework needs to be fleshed out and made a reality. Sometimes, that can be just as challenging as getting the approvals and funding. Hiring a few critical people, forming a data governance council and getting the right people to participate, gaining a few early wins, resisting the temptation to make everything a data governance problem - these are some of the issues you'll deal with in the early stages of your program. Try to imagine how you'll confront these questions ahead of time. Nothing is insurmountable, but be prepared for growing pains as your program gets off the ground and begins to stretch its political legs.
Ultimately, data governance is a political challenge. It needs to be led by a person with a lot of business savvy and political skill. Budgets, headcounts, business processes and organizational alignment will be more important than technology selection and implementation. But in the end, there's not much point in trying to manage your data on an enterprise-wide basis if you're not going to govern it that way too.