The Future of Big Data in the Mortgage Industry

mortgage industry

It’s common knowledge that Big Data has arrived in the Mortgage Industry. One of the most important questions leaders in our industry need to ask themselves, of course, is “Where is it all going?” We’re going to give you our take on this issue in just a moment. But first, let me give a short synopsis of what Big Data is for those who are new to this field.

Historically, all the data computers used was set up in highly structured data bases. In other words, we had separate fields for each piece of data and we spent a lot of time and effort to make sure all the data was clean and accurate. Big Data does away with that. Big Data reads data that was never meant to be analyzed by a computer. This includes everything from Tweets and Facebook postings to newspaper clippings. All of these were written for human consumption, not for computer processing.

Big Data cut through that. Big Data is able to read all of this unstructured, messy stuff that was never meant for computers and then makes sense of it. It other words, it can read Tweets and Facebook postings and data from hundreds of different sources that are written in incompatible styles and assign meaning to what it’s reading. In the mortgage industry, this means that we can now tap into huge reservoirs of information that were always available to us before – data that is in the public domain, but we could never get a computer to work with it.

Now let’s take a look at where Big Data is going to take the mortgage industry.

Big Data in the Mortgage Industry

One important Big Data application is pre-populating mortgage applications. In other words, Big Data will mine data from bank records, publicly available data bases, social media sites, and other sites to collect all or nearly all the information required for a mortgage application. This will leave the applicant with the option of either clicking to ratify the pre-populated application as accurate or, on the other hand, edit a few fields here and there to fine tune the application.

Another approach here is for prospective home owners to complete their mortgage applications as they always have and then the mortgage company’s computers will compare the pre-populated versions with the applicants’ versions to identify discrepancies.

In either case, the objective of this exercise will be to enhance the accuracy of the data in the applications at the same time the system reduces the burden on the applicants.

We can also see that computer algorithms will score mortgage applications using machine learning algorithms. These algorithms will approve or deny the applications immediately. Approved applications may be forwarded for processing right away. Rejected applications will qualify for a human review if the applicants don’t feel they have been scored properly. The goal of this instant evaluation will be to eliminate the delays in the current manual evaluation process – delays that are often measured in weeks.

We can see that Big Data will be instrumental in projecting the number of applications for new mortgages or refinanced mortgages in specific geographies and specific time frames. Further, Big Data will project the total value of these mortgages. These projections will help mortgage companies reposition their people and processing power based on projected market demand. These projections would be based on the current mortgage portfolio the industry has in place in various geographic areas coupled with scenarios about shifts in mortgage interest rates.

Spending on Big Data applications and technology will soar. In 2014, 2015, and 2017, we’ve seen Big Data spending in the mortgage industry at $2.6 billion, $2.8 billion, and $3.2 billion respectively. We are going to see spending on Big Data continue to climb as the number of success stories grows.

The mortgage industry is going to suffer a severe shortage of Big Data analysts who know how to manipulate the huge and ever-growing quantities of data that will become available. We are going to need professionals who can manage the enquiries in ways that lead to highly defensible conclusions. The growth in the demand for Big Data analysts is going to outstrip the supply.

We are going to see the rapid growth of specialized firms that assist mortgage companies plan for and implement Big Data projects. This function is going to outsourced rather than treated as a core competence for several reasons. First, most mortgage companies will find it far too expensive to build their own in-house facilities. Second, the process of building their in-house facilities will take too long and are liable to face many dead-end alleys. Third, they will not be able to attract the talent they need at a price they can afford. Fourth, the management in existing mortgage companies will need to go through a steep learning curve that is best handled by a specialized firm. Over time, we can expect mortgage companies to build teams of in-house Big Data talent while leaving the technologies to cloud-based firms.

The privacy issue is going to become a big issue in Big Data. Although everything Big Data practitioners do is legal, the act of mining social media on a wholesale basis was never considered when social media sites were first introduced. We are going to see some interesting and instructive debates on ethical issues over the next decade before we see a consensus emerge. Any legislation passed before those ethical debates come to closure will prove to be ill-conceived and counterproductive.

Decades ago the local bank manager knew his customers well and was in a position to make an informed judgment call about the amount of credit to be extended. Bank managers rarely make those decisions in retail bank branches and mortgage companies today. Rather, those decisions are made by a committee – often in another city. We need to reinvest some humanity into the decision-making process. Incorporating social media will go a long way in that direction.

The mortgage approval process is going to become more transparent. At the moment, borrowers only know whether they are approved or rejected, but they rarely have an idea why they were slotted where they were. In the future, mortgage companies will be in a position to coach their applicants very specifically about what they need to do to be approved.

We are going to see many new, non-traditional players in the mortgage industry. They will spring from places like Silicon Valley. They will offer better service at lower costs than banks and traditional mortgage companies. For example, the Lending Club facilitated $3.6 billion in loans in the first six months of 2015. Likewise, Prosper is growing fast.

We can expect the Federal Housing Administration to develop a growing number of regulations that the mortgage industry must comply with. Many of these regulations will apply to a company’s portfolio of mortgages rather than any given mortgage. Mortgage processors will continue to ensure that they comply with application specific compliance issues, but they cannot be expected to deal with portfolio-wide compliance issues. In fact, it is unlikely that it is humanly possible to do so. This means that mortgage companies will necessarily embrace Big Data to do that job for them. Failure to do so means that they will face stiff penalties in court. It is far better for these companies to catch non-compliance failures on their own and take action than to face their regulators in court.

Carl Pry, a managing director at Treliant Risk Advisors, said “It’s in every bank’s best interest to get one step ahead of the regulators and understand what that regulator is going to know and find. They need to resolve any discrepancies [and] do any file review analysis needed to be able to explain any disparities before the regulators find them.”

Big Data is going to help reduce the risk in mortgage lending. Big Data will help brokers advise their clients about school performance and community crime rates. This will help the buyers make better informed decisions and, ideally, lead to lower risk mortgages.

Big Data is going to prove instrumental in flagging potential fraudulent mortgage transactions. The FBI and other law enforcement agencies are developing increasingly sophisticated techniques to identify potential abuses. Big Data algorithms will incorporate these fraud detection techniques into their algorithms and trigger pre-emptive enquiries.

Residential property appraisals will become more objective and more accurate. Big Data will propose the most appropriate neighborhood comparables. It will develop appraisals using industry standards that will be driven by an algorithm. MReport claimed that, “More than 30 percent of loans fall short of the collateral valuation agreed to between customer and loan officer.” Big Data will help fix that.

New regulations and compliance issues are making the appraisal process increasingly difficult. That, coupled with the fact that the number of qualified appraisers is not keeping up with the demand, means that the industry must necessarily rely on broad based, sophisticated tools like Big Data. This trend will continue.

We can expect federal compliance regulations in the mortgage industry to be applied ever more strictly. In the last few years we’ve seen fines and settlement agreements that were even more onerous that Dodd-Frank required. In these cases, the government targeted the big boys like Wells Fargo Bank, Bank of America, CitiBank, PNC Bank, EverBank, JP Morgan Chase, One West, Santander Bank, and U.S. National Bank. Given the government’s practice of starting with the big companies and working down to smaller companies, it is not hard to see what is liable to happen.

We can expect to see Big Data analysis of non-monetary defaults on mortgages to become more common if not universal. Here, I’m talking about flagging accounts where payments were made early and with an extra principal payment to being made on time with no extra payment. Or we will find homeowners whose home owner association is suing them. Or maybe the local government put a lien on the property on the grounds that the property is uninhabitable. Or the couple is getting divorced. These are all early warning signs that Big Data will track as a matter of course.

Small mortgage companies that cannot afford to buy the necessary technologies will be squeezed out of business. Larger companies will buy them.

Mortgage companies are going to increasingly focus on building higher quality portfolios with fewer staff. The only way to have a smaller staff complement and a larger mortgage portfolio is through automation. That should be obvious. Automation in general and Big Data in particular is the way of the future.

Just to wrap up, I want to make it clear that Big Data is already having an impact on how the mortgage industry operates and we are still at the early stages. We are going to be in for a very interesting ride over the next few years.

Big Data Case Studies in Education

big data case studies

Big Data Case Studies with Proven Results

Big Data Case Studies: Coursera

Coursera provides education from leading universities around the world delivered over the internet. The instruction is handled through data streaming videos. Coursera tracks how its students watch those courses. Students might “rewind” to watch a section a second time. Or they might fast forward – skipping stuff they think they already know. Or they might go over the same course several times. Or they might just quit and walk away. Whatever they do, Coursera tracks it on a student-by-student basis. The company learns from this experience. It learns what works and what doesn’t. Occasionally it throws in a pop-quiz to see how well the students are learning. But there’s another reason, too. The company wants to see how well it’s doing. It’s a kind of self-evaluation. When the course designers realize that the learning process is not going as they had expected, they can go back and rework their material based on real-world feedback.

Big Data Case Studies: Arizona State University

Arizona State University, like many universities across the country, has its fair share of freshman students who are genuinely challenged in mathematics. One third of their freshman classes earned less than a C in math. Interestingly, this one score has been a reliable indicator of whether students would eventually graduate and collect their degrees – or drop out. To deal with this, ASU worked with Knewton apply its adaptive learning techniques. In just two years – from 2009 to 2011 — the pass rate in this course jumped from 64% to 75% at the same time the dropout rates fell by 50%.

Big Data Case Studies: West Virginia University

Simon Diaz, a professor at West Virginia University, was very curious why so many students who enrolled in online classes dropped out. One of the key rationales for providing online classes with streaming video at times convenient for the students was that the students wouldn’t feel shackled to a schedule that was incompatible with the daily realities of their lives. Using Big Data analytics, he looked at 33 variables for more than one million students. These variables included everything you would expect like age and gender to things you wouldn’t expect like military service and class size. What he discovered had never been obvious to anyone else before. The more classes students took at any one time, the more likely they were to drop out. Simply by reducing the number of courses students enrolled in at any one time would increase retention rates. But financial grants to students require those students to take a minimum number of courses. In other words, public policy was at odds with good educational practice – a conundrum that no one had discovered before based on a policy that had probably never been thought through with any empirical evidence. Another win for Big Data in Education.

Big Data Case Studies: Kent State

Kent State uses analytics to track student activity and project the likelihood of success. It tracks students over a ten-year time period collecting data about their majors, classes, demographics and other factors. Their system highlights the students at risk with red, yellow, green indicators. The reports help advisors focus their efforts on problem areas. Steven Antalvari, Kent State’s director of academic engagement and degree completion, said, “This data has helped us peel away certain layers faster, allowing us to spend the bulk of our time together working on the student’s purpose, goals, and career development.”

The Top Players in the Education Technology Industry

education technology

Here are some of the well-recognized in the Education Technology (EdTech) sector – in no particular order. These are the companies that are developing the paradigms that will shape Big Data in Education. They are also the companies that are developing the technologies to implement those paradigms and offer them to educational institutions.

This is important because, until recently, schools have not needed to look outside their own walls for the tools they needed to do their work. The obvious exceptions are textbooks and, starting some 60 years ago, general purpose computers.

Educase Conference – This company is growing fast. It offers systems to store Big Data in a cloud and perform analytics on that data to make sense of it for administrators.

SAS is a well-established company that dominates the advanced analytics industry with almost 32% of the market.

Renaissance Learning A few years ago this company was sold for $1.1 billion. Renaissance is a testing and student data company. At the time of its sale, it had data on the test results for 10.1 million school age children.

InBloom This company is a middleman between school districts and education technology companies. It handles the data storage and distribution of student data to authorized users. The Bill and Melinda gates Foundation and the Carnegie Corporation of New York had so much confidence in this venture that they kicked in $100 million. However, social concerns about data security grew to a fever pitch and the company withdrew its offerings.

Coursera Coursera is a start-up company that offers courses over the Internet. It offers accredited courses from Illinois, the University of Pennsylvania, Johns Hopkins University, the University of Michigan, Stanford University, UC San Diego, Duke University and 150 other universities around the globe. Students can even earn a master’s degree in business, accounting, data science, and entrepreneurship through Coursera.

Noodle Noodle is based in New York. The company offers fact-based information to help prospective students choose an elementary school, a graduate school, a summer camp, or even a tutor.

Knewton – According to its website, Knewton recognizes the value of adaptive instruction and education technology: “Individual students bring different skills and different challenges into the same classroom. Knewton’s pioneering approach to adaptive learning draws on each student’s own history, how other students like them learn, and decades of research into how people learn to improve future learning experiences.

Big Data in Education: Full of Promise, Uncertain Future

big data in education

This is What Big Data in Education Looks Like

One educational practitioner used Big Data to catch an anomaly in a course that was designed to progress smoothly from one module to the next. He found that the students in the class progressed from module 1 to module 7 as expected. At that point, however, most of them went back and replayed module 3 again. It became very clear that the material in module 3 hadn’t “stuck.” This led the course developers to revisit that module and upgrade it. They did this even though none of the students complained about that module. By monitoring what students actually did on a massive scale, the company saw an opportunity to upgrade its course and did it. In this blog, we will discuss the importance of big data in education.

In another instance, students were stumbling on a particular question and were notified immediately that they missed it. Many of those students read the related forum material, reworked the quiz, and got the answer right. When the course instructor discovered this through Big Data analysis, he inserted a recommendation in the course for students who got the answer to that question wrong: He referred them to the forum post that had proven useful for everyone else.

Spanish speaking students studying English via Duolingo would stumble and fall when learning the English pronouns he, she, and it. This led to high dropout rates. Why? Well, Spanish doesn’t have an equivalent to it. This was a new concept – and new way of thinking – for Spanish speaking students. The solution was simple. The course postponed the introduction of the word it for a few weeks and student retention soared.

New York City has a program School of One. In this school, each student gets his own playlist of modules to study. The students need to learn math. They go at their own speed. If one module doesn’t do the trick for them, they try another. Now, the real question is, “Does it work?” Well, independent studies by a private educational service reported that students who went through this program did substantially better than those who did not. Yes, it works.

What is Big Data in Education?

There are two major areas of interest in the field of Big Data in Education: institutional and educational.

Institutions collect masses of data from traditional sources as well as new sources to develop their policies and plans. The new sources include Facebook posts and Twitter tweets to get a sense of the sentiments among current students, prospective students, and the community at large. The institution can also pick up macroeconomic and microeconomic data that are useful but were prohibitively expensive to include before.

Educational or instructive purposes are intended to personalize the learning process for each individual student. Here, schools at all levels can collect detailed data about students’ progress through their learning journey on a moment by moment basis. The idea is that the system can identify when a student is caught in a vortex that prevents her from making progress. At that point, the system could notify the teacher about the problem a student is having at the moment it occurs. On the other hand, the system could be designed to introduce a tutorial that deals with the problem area as it occurs, not weeks later when a failing score highlights the students’ learning problems.

This student oriented real-time instructional intervention has several benefits. First, it helps the students well before frustration, disillusionment, and failure set in. This helps the students to become proficient in the subject material – even master it. It also has the benefit of assisting the teacher to focus her attention on just the sort of help that is needed. This is particularly beneficial in large classrooms. Students’ success in the classroom will lead to them staying in school and gravitate toward matriculation. Success in school is correlated with success in the work place.

In addition, as students stay in school and graduate, the school builds its reputation as a place where students can come to succeed. This attracts new students. Further, by keeping students in school until graduation, the school improves its revenues as well as its reputation. These retention and graduation rates loom large in school evaluations.

The US is Not a Leader

Ironically, the US is not a leader in education when compared with other developed countries in the world. We have seen class sizes in the public schools grow to the point that teachers can no longer provide individual attention. Funding to schools at all levels is constantly being cut back. There is no question that America is home to some of the leading universities in the world, but those universities are not characteristic of the country as a whole.

One way of dealing with this growing gap between the quality of education and access to education in the US compared to other countries is to adopt distant education and Big Data technologies. These technologies promise to offer education aligned with the students’ schedules, not the class room schedules. Further, it promises to offer meaningful tutorials on problem areas tailored to each student as and where they are needed. These benefits are likely to be compelling in informing educational policy.

AltSchool May Be the Extreme

If you want to know what Big Data in Education on steroids looks like, look at AltSchool. This San Francisco Bay Area company will record everything about their students while they are in school. That means EVERYTHING. It will track how they go through their learning experiences – heart rate, eye movement, facial expressions, movement from one part of a computer screen another, how long their mouses hover over items on their screens. They will record every word. Almost every thought. Everything. All of this data is then fed into a Big Data database. Top notch data scientists will comb through this data to learn in detail how to personalize the learning experience for each child.

AltSchool might find that some students improve their mathematics studies after exercising in the schoolyard. Or a student starts incorporating new words in her vocabulary after watching a particular video. Then the school will incorporate those insights into the student’s daily routine and see whether the benefits persist over time. In fact, the school planners would use Big Data to look for an ongoing series of tweaks they could make. That will provide a stream of changes that may (or may not) provide enduring value. This is personalization at the extreme.

Some argue that this sort of super tight oversight smacks of Big Brother — and maybe it does. But if it pays off in terms of enhanced results for the students, then it may be worthwhile. There are probably hundreds of practices we respect in everyday life today that may have seemed strange – even objectionable – our forefathers a few generations ago. For example, it was less than three generations ago that it was common practice to whip children who performed poorly in school; today that practice would be unheard of.

Huge Investments

GSV Advisors estimated that the e-learning market in the US is over $100 billion. Further, it’s growing at 25% a year. Well-established companies like McGraw-Hill, News Corp., Pearson, and Kaplan have spent billions to get into this market. Further, there are a lot of start-up companies mushrooming in this space as well. We’ve listed just a handful of some of the notables below, but there are many other worthy companies that didn’t make this list

Digital Transformation Technologies Help Increase Revenue

Man using Ipad to look at the digital transformation

Digital Transformation

Digital Transformation technologies are one of the hottest trends in the private sector today. In fact, it will spell the difference between the winners and the losers. Companies that get it right are seeing their market caps soar; those that get it wrong are growing slowly or – even worse – falling by the wayside.

The key to establishing and running a successful business is adopting Digital Transformation technologies – including mobile apps, digital marketing, and predictive maintenance, to name but a few. It is not enough, however, to simple introduce these technologies, businesses should leverage these technologies to their advantage.

To truly leverage such technologies, business should change the organizational structure and hiring and training methods to complement these new digital technologies. Studies have shown that using these transformative technologies is now the greatest predictor of a business’s revenue growth. In today’s digitally connected world, adopting these technologies is the fundamental factor in determining whether or not a business will succeed or fail.

To learn what Digital Transformation will provide the greatest payoff for your business and just how to leverage these technologies for your advantage, download our guide to increasing your business revenue through Digital Transformations.

To Download the Guide Click on the Image Above or Click HERE

Read more about the top digital transformation technologies today.