Cleaning Up Your Data

It’s no secret that many lenders, investors, and servicers are bogged down with operational inefficiencies and rising costs that threaten their existence. At the MBA Annual Conference, Bill Emerson, Quicken Loans’ CEO, summed up this challenge in one statement: "The industry has lagged behind [other sectors] for a long time. We've been in the data business for a long time, and we're still a paper-intensive industry.”

While mortgage companies are well aware of the challenges, what has been elusive until now is a solution that enables them to streamline mortgage operations without massive, costly infrastructure projects that could span years. Fortunately, emerging technology can drive significant process improvements across the enterprise without massive investments of time and resources.

Companies are moving away from labor-intensive processes by adapting technology that makes data easier to work with—but true innovation cannot happen without clean data, or a “single source of truth” to power all their business applications with confidence.

Clean data is critical because most legacy systems struggle with ingesting data from multiple sources. Information is often held in multiple, siloed systems that make it difficult for staff to extract what they need for critical business functions and insights. Employees end up spending time manually turning documents into digital records, then checking and re-checking data to avoid errors, which results in a loss, damaged reputation, and even regulatory violations.

For originators, the staff has to quickly take in a large batch of documents and ensure data accuracy before the loan can be underwritten. To complicate matters, each mortgage loan folder typically has 2.2 versions of every document. Documents like loan applications or bank statements can have up to five versions. Throughout the process, the underwriter, closer, and post-closer are checking the data again before performing their actual work.

For correspondent lenders, a single submission from a customer could involve 500 pages of documents. Before a file can be evaluated for its risk profile, lenders must adhere to a checklist to ensure all the right documents have been received, and go through a grueling “stare and compare” process to ensure the data is accurate. Lastly, employees have to manually enter data into the system of record.

In servicing, similar issues exist with onboarding tens of thousands of loans and validating the data on the mortgage servicing rights (MSR) data tape against the documents. Servicers have to race against the clock as audits need to be completed quickly so the Notice of Transfer can be sent to the consumers within the required 15 days. Auditing 100 percent of the loans in a compressed time frame creates significant operational challenges for servicers.

All these problems can be solved through software-powered automation that makes documents and data easier to work with. It starts with a cloud-based mortgage data store that absorbs millions of data points from documents and digital sources and standardizes them into an authoritative record for each loan that can be traced back to its source. Then intelligent tools can be deployed to work alongside existing systems to make that data accessible for better and faster decisions throughout the mortgage life cycle.

For example, technology can help companies say goodbye to “stare and compare” by providing a single place to see and compare data across all sources, including the Loan Origination System (LOS) and supporting documents. Mismatches are automatically flagged so staff only need to handle exception review. Companies investing in such emerging technologies are seeing serious operational efficiencies that speed up cycle times and reduce risk and costs.

Clean data to drive workflow automation is just the beginning. The data can be used for analytics to support business decisions, provide full transparency into the assets sold in the secondary market, and ultimately help the mortgage industry avoid another big short like 2008.