Data analytics, the statistical analysis of large quantities of data, is nothing new. Some would argue that scientists and statisticians have been analyzing data sets to uncover trends for hundreds, perhaps thousands of years (depending on how “analysis” is defined).
However, the way it’s being used today, with computers able to help assist data scientists in the analysis of massive amounts of information, really only began well into the second half of the 20th century. The knowledge provided by data analytics has transformed countless industries, particularly when it comes to the ability to predict behaviors and outcomes, a subset of data analytics called “predictive analytics.”
From Descriptive to Predictive Analysis
Whereas descriptive data analytics collects and aggregates data to examine the past and the present, predictive analytics takes it a step farther– using machine learning and focused algorithms to predict future behaviors, attitudes, and trends. Through data analytics, information is analyzed with greater speed and efficiency. With predictive analytics, the analyzed information helps businesses make decisions with greater transparency into future business outcomes.
Businesses are clamoring to integrate predictive analytics into their operations because it makes processes run better.
Many businesses, for instance, are finding that data forecasting can streamline business operations, aid in reducing risk, and, ultimately, increase revenue with more speed and efficiency than ever before.
Just take a look at some case studies:
Predictive Analytics in Retail & E-Commerce
Predictive analytics is being used within both brick and mortar and e-commerce retail to analyze customer buying patterns and other behaviors. With this information, stores are able to:
- Provide personalized, targeted marketing to consumers
- Identify effective advertising outlets
- Anticipate traffic by location and time period
- Note customer pain points to increase retention and build loyalty
- Improve store layout or UE by tracking customer movement
- Effectively determine price points and inventory needs
Case Study: Amazon
Amazon is shipping customers orders before they’ve been placed. Their newly patented “Anticipatory Shipping” uses a predictive analytics model created using customer data such as prior purchases, order frequency, cart contents, and search history. The algorithm ensures the relevant products are shipped to the closest hub, in anticipation of the customer order. The result: Customers will enjoy the benefits of faster shipping times at a lower cost, increasing their brand loyalty and providing Amazon with an increase in sales.
Predictive Analytics in Banking and Lending
Like retail, lenders are using predictive analytics to improve customer experience– analyzing current buying patterns and other customer data to predict customer pain points and identify potential product offerings. What’s even more exciting is that some businesses are also using predictive analytics to safeguard their customers’ information and assets.
Case Study: Capital One Bank
As the Director of Data Science at Capital One Bank, our own Chief Data Science Officer, Andy Mahdavi, was on the team that developed the bank’s transaction monitoring system. Using behavioral data as a benchmark, the algorithm identifies activity that falls outside the customer’s behavior profile and alerts the customer in real time. As a result, the bank is able to deny fraudulent charges immediately and with higher accuracy, setting in motion an investigation that protects the customer’s money without causing undue alarm. At the same time, compared to the system previously in place, Capital One Bank materially lowers its fraud losses each year.
Predictive Analytics in Sports
Before Facebook, Google, and Amazon, there was Moneyball. The movie made famous by Brad Pitt revealed how Billy Beane, the General Manager of the Oakland Athletics, used “sabermetrics” to guide draft picks. Now, predictive analytics is a major tool across all aspects of sports management.
Case Study: 2018 Google Cloud and the NCAA example
In 2018 Google Cloud teamed up with the NCAA to create predictive models that analyzed over 80 years worth of data pertaining to team performance. The goal: to be able to predict a team’s success in real-time.
The experiment took place during the Final Four games in San Antonio. During the first half of the game, data scientists collected information about the teams’ performance and then combined this data with historical data to predict performance during the second half of the game. The results weren’t perfect, but they were pretty impressive. Here are some of the results from Michigan versus Loyola Chicago game:
- Prediction: 37 three-point attempts
- Final tally: 38 three-point attempts
- Prediction: 29 rebounds
- Final tally: 29 rebounds
Predictive Analytics in Healthcare
Perhaps the most exciting use of predictive analytics is in healthcare. As more medical facilities adopt the use of EHRs (electronic health records), more data is available to support algorithms that can help providers better care for their patients. In the healthcare sector, predictive analytics is helping to save lives.
UC Davis has been collecting EHR data to create an algorithm that can detect sepsis in admitted patients before the patient exhibits symptoms. Sepsis is life-threatening, and oftentimes, when a patient begins to experience symptoms, it’s too late.
Using medical data, analysts working with Kaiser Permanente created a risk score to predict the likelihood of dementia in diabetic patients, enabling at-risk patients to be proactive about treatment.
A study conducted by the Mental Health Research Network in conjunction with Kaiser Permanente researchers concluded that an analytics model using EHR data and patient questionnaires were far more accurate at predicting suicide risk than former models using fewer data points.
Predictive Analytics in Human Resources
Every business leader desires a high-performing, loyal workforce. Yet outdated hiring methods that are dependent on human-guided decision-making are subject to bias and can be highly inaccurate. With predictive analytics, human resources is no longer subjective.
Case Study: Evolv, Inc.
Before starting Doma, Max Simkoff was a co-founder and CEO of Evolv, Inc., an enterprise software company that used predictive analytics to provide solutions for workforce optimization. Through the analysis of employee data, the Evolv product algorithm could help companies predict how well a candidate would perform if hired and how well current employees would perform if promoted.
And Now, Predictive Analytics in the Mortgage Industry
Predictive analytics is all about improvement. Helping businesses improve operations, management, and customer service with greater speed and efficiency. It’s this understanding of the revolutionary nature of predictive analytics and machine learning that has led Doma to develop predictive underwriting. Now with the help of Doma, a clear to close can be issued in a matter of minutes instead of days or weeks. Not only does the Doma solution increase the speed and reduce costs. But by surfacing issues sooner in the process, it’s also been found to increase the certainty that an opened file will close.
Request a demo to learn more about how Doma can support and transform your mortgage lending solutions.