If technology is the driver, then the management of information must be the outcome. Jill Dyché is vice-president of SAS Best Practices, and an author of three books on the business value of technology. She is a frequent speaker and blogger, and her most recent book is Customer Data Integration: Reaching a Single Version of the Truth. She spoke with editor emeritus Paul Crookall.
What are the data management issues at the forefront of your work with governments?
I’m hearing two main questions. One, “How do we bring people together around data?” In the public sector the drivers are largely the regulatory issues around sharing, the creation of policies around data, and the understanding of how to execute them. Executives are also interested in the organizational implications of interagency data sharing.
The second question is, “How do we better achieve our mission using big data?” Executives on both the business and IT sides want to understand how big data and analytics can help their agencies provide better service.
Would one example be Canada using existing data to automatically enrol citizens in the Old Age Security Program, rather than have them complete a paper form that contains information government departments already have?
That’s a good one. There are lots of examples of operational efficiencies driven by analytics. Managers want to understand not only the strategic promise; they want to know the tactical next steps. And with big data, this includes how to deploy unstructured data, like social media interactions.
Do you find that CIOs are increasingly being asked to take a leadership role in change management?
Yes, but it is in proportion to the vision of the deputy minister seeing the business promise of IT. CIOs for too long were seen as responsible for “keeping the lights on.” They’ve been measured on things like system “up time,” rather than business value. Increasingly, though, they’re being asked to provide strategic insights, and their innovation potential is being recognized.
We’re seeing two changes in support of that. More CIOs are reporting directly to the CEO or deputy minister, rather than through the chief operating officer or ADM Finance. And the average tenure of a CIO is increasing as they become more familiar with business strategy.
As a non-IT manager, what do I need to know about big data, data management and data analytics?
First, now is the time to start adopting analytics and big data solutions, if you haven’t already. Costs have come down, the enabling technologies are cheaper and better, there is a lower barrier to entry than ever before.
Second, you should define the need, the pain point, or the problem that data would help you solve. For example, data analysis at a major postal service showed a big problem was fraud and theft of the mail. This not only affected profitability but customer trust.
Third, consider what new processes you want to add. Maybe new customer data can help your agency refine outreach campaigns or improve financial reporting, and that means enlisting different stakeholders or adopting new information deployment methods.
Fourth, leverage the data. For example, in Michigan, they had a problem with deadbeat dads, who weren’t paying their child support. The state passed a law that enabled agencies to share data, so now the agency managing child support payments can connect with other government departments, including the Department of Motor Vehicles to prevent licence renewals until payments are up to date. This led to both an increase in payments, and a corresponding decrease in public aid. So in a sense the program stimulated both cost savings and tax revenues.
As an IT specialist or general manager, how do I drive innovation and foster a culture where data is used to drive value into the organization when my staff are demoralized from budget cuts and feeling a loss of control?
The “eureka” phenomenon doesn’t just happen. Leaders need to support more formal analytics that can help better achieve the mission while saving money. Show an early “win.” For example, one agency we worked with used data visualization tools to pinpoint geographic “hot spots” for food stamp fraud, saving money for the program. This had both “hard” and “soft” benefits, and the ROI was critical in ongoing funding for the program.
Your article in Harvard Business Review describes “knowledge discovery.” What is that, and why is it new?
Traditional database inquiry requires some level of hypothesis, but mining big data reveals relationships and patterns that we didn’t even know we were looking for. It is mining data with no starting hypothesis, to see what might be there. Software tools find the patterns and tell the analyst or data scientist what, and where, they are.
In health care, for example, Stanford University researchers were looking for trends in breast cancer cells, expecting to see trends in cell proliferation rates. But they also discovered that surrounding non-cancerous cells contribute to cancer cell growth. This technique is used less in government; it will likely come later, after trust in analytics is established. Once there’s a critical mass of detailed data, data analysts can then start looking for natural patterns.
Comstat seems to be an early example, where police, fire, welfare, housing inspectors, and other diverse departments shared data, numbers, and maps to identify problem areas, connect incidents, and share tactics and responses. It helped drive down crime (and fires and derelict housing) in New York, Baltimore, and other cities where it was used.
Yes, there are great big data case studies in the law enforcement arena. Tax and health care agencies are also early adopters of big data in government. Data is extensive in these areas, and we need to find ways to share and analyze it across jurisdictions. And that’s where not only government agencies but citizens themselves will see the value of analytics.
Jill Dyche will be a keynote speaker at CGE’s Leadership Summit in Ottawa April 4. See http://cgeleadershipsummit.ca/