Jenny Schultz has been with Freddie Mac since 1998. We have tried to do the “data thing” more than a few times, but it never worked because the business didn’t see the value in using their data and managing and governing it. This is what she said at the DataVERSITY® Enterprise Data World Conference: “Freddie Mac Single-Family Business: A Case Study in How to Operationalize a Data Stewardship Model.” She shared the stage with Van Lin and Stephanie Grimes.
After four or five tries, a recent project finally caught on because, says Schultz, the business was behind it. The thing that made the Data Stewardship Model work, she said, was getting the right people on board.
Schultz and co-presenter Van Lin spoke at length about how they operationalized a new Data Stewardship Model at Freddie Mac. Stephanie Grimes is the manager of data governance for the single-family line of business. Both of them work with her to make sure that data is safe. The presentation was broken up into three parts: (1) Ready, setting the groundwork and setting up the programme; (2) Set, finding and working with stewards and stakeholders; and (3) Go, the programme itself. However, this introduction only talks about the “Ready” part of the operationalization process: the steps from planning to execution.
It’s what Freddie Mac does
Freddie Mac’s single-family division gives money to banks so they can lend money to people who want to buy single-family homes. It has helped people get 62 million loans since 1970. All of the loans we fund have 75 different types of data that we keep track of. When it comes to single-family home loans alone, Schultz said, “We have borrower information for 114 million people and 94 million properties.” Not all of those businesses are included. Servicing, their capital markets business, and their multifamily business are not included. “The short storey is that we have to deal with a lot of information.” We only do that.
Freddie Mac and Fannie Mae are both government-owned programmes that help banks lend money for mortgages. Schultz: “We live in Virginia, and they live in D.C., so we call them our friends across the river.” To be able to compete with them and respond to new programmes and products that they put out in the market, we need to be ready. Freddie Mac’s mainframe-based infrastructure is very complicated, which makes it hard for the company to react quickly to business problems. In the end, they said, “We need to make our architecture simpler and have a north star: where we want to get to.”
When you start making a programme, you have to figure out where the pain points are.
The head of the single-family business asked the CPA to find a way to use their data and make the business more valuable so they could compete better. The CPA, who isn’t a data person, decided to talk to important people in the company to see if there was a need for a solution. Schultz: “We went on what we call the “listening tour,” and that’s what we did. They talked to their stakeholders, coworkers, the internal audit department, compliance, and the privacy office, to figure out what problems they were having and what business problems they needed help with, so they could figure out what help they needed. She thinks this is one of the most important parts of the program’s long-term success. When you start a new project or change something, talk to your business stakeholders. We can’t just think we know what the business needs.
It was after the listening tour that they decided to work with the Enterprise Data Management Council (EDMCouncil) and use their Data Management Capability Assessment Model (DCAM) to make a model from people’s ideas. The DCAM helped them make a plan that would get them from where they were to where they wanted to be: a well-organized data set, which was the goal.
Setting data standards is the best way to do this.
Instead of writing data standards in a room with her team, Schultz said they decided to set up working groups to make sure everyone was involved. This process could have been done in a few weeks. They wanted standards that were true to the world that people worked in. Do they have any problems? What solutions can we come up with to help them? People are working on the fifth version now. The first draught took at least 90 days to put together, and they’re working on it now. It’s also because we listen to people and make changes based on what they say.
Different types of data need to be handled in different ways, so the groups made the Data Governance standards match the data’s use. This is what she said: Financial data, mission-critical data, and data for R&D all have different risks and don’t all need to be controlled the same way.
Standards also set out roles and responsibilities, which included an executive committee that looked after high-level priorities and made sure everyone agreed on road maps. Schultz’s supervisor added more oversight to make sure that the planning kept going. If necessary, she had the power to overrule the process and make a final decision.
How do you make sure that your data is safe?
Our data-driven company has no employees who don’t play a role in Data Governance, Schultz told the group. Owners of data must give metadata. Owners and stewards of the data need to know what users are using the data for and what their quality thresholds are so that they can figure out who to talk to if they want to make changes in the data. Data is not a technology problem, but rather a “people-culture change-communications problem,” Schultz said. This has led her to believe that data is not a technology problem, but rather a “people-culture change-communications problem.” If we didn’t have to deal with all the people, “We’d be done with data by now.”
If you are a data manager, you have to do some “data therapy.” Schulz calls this part of her job “data therapy.” Often, people just need to get their thoughts out of their head. I listen all the time. Many times, it’s just about feeling like you’re being heard. “No, it’s not about that.” A big part of what makes a big difference is how people interact with each other and connect with each other.
Source: data science course malaysia , data science in malaysia
Leave a Reply