Predicting buying behavior in retail
With the retail industry seeing nearly $ 4 trillion in sales annually, it’s no wonder why companies like Amazon and Walmart regularly use analytics to learn everything they can about their customers.
For example, in 2004, Walmart mined transaction data in its stores to understand buying habits at certain times. They found that just before the hurricanes hit, sales of strawberry Pop-Tart increased seven times along with beer. Of course, Walmart used this as an opportunity to store its shelves. We discuss the technique they used in our introductory guide on data mining.
Amazon has previously used predictive analytics to create personalized product recommendations based on buying patterns.
Most recently, Amazon is seeking to use predictive analytics for predictive shipping. In other words, shipping products to customers before they even buy them based on their behavior on the Amazon platform. This can lead to uncannily fast delivery times.
What if I’m not a business?
Predictive analytics is not reserved for the big players. Many of today’s retail POS software are good at collecting customer data and integrating with other systems such as CRM, supply chain and inventory management to be used for predictive analytics.
Successful resellers are able to collect and combine data from all touch points, such as e-commerce sites, mobile apps, store placements, social media platforms and more. Analyzing this data will help you understand your customers on a deeper level and predict their behavior in a more personal way.
2. Detection of disease in the health care system
There are more than 36 million patients in U.S. hospitals alone; you can only imagine how much health data it is about.
But the healthcare industry is not as much focused on the consumer journey as it is focused on analyzing data to improve diagnoses and predict outcomes based on certain health factors. Interestingly, Jeff Howell, director of growth at AlayaCare, gave us a real example of how they used predictive analytics to investigate adverse health events in seniors.
“We worked with Element AI to produce an algorithm that successfully predicted adverse health events in seniors (in their homes). Seniors would take a number of vital things every day (for example, a blue tooth weight for weight). The algorithm digested the vital ones. and combined it with clients’ ICD-10 diagnosis, age and gender. We have successfully reduced hospital admissions and emergency room visits by 73% and 64% among a chronically ill patient set. ”
These visits are extremely expensive for any healthcare system. This external patient monitoring software is linked to patients’ operating software for their home health authorities, so when the health risk score becomes too high, the home health authority can intervene with a visit to get the client’s health back on track. “
Curators of content in entertainment
The entertainment industry, more specifically digital entertainment, benefits greatly from the use of predictive analytics. Let’s look at some of the ways in which today’s digital media and entertainment giants utilize big data to shape viewing experiences.
We know there are more than 100 million active Netflix accounts today, equivalent to billions of hours of streaming digital content. All of this data helps Netflix build predictive models to keep their consumers happy and expose them to relevant shows.
So what are some types of data Netflix uses for their models and algorithms? Some of the user data includes:
The preferred genre of content.
Search keywords when searching for content.
Ratings.
The preferred device for viewing content.
Dates seen, and in some cases revisited.
Time spent viewing content previews.
When the content is paused and at what time.
These metrics and many more are important to the success of entertainment streaming services. In fact, Netflix used this data to make its show House of Cards, claiming that they already knew it would be a success based on the results of predictive data analytics.
4. Prediction of maintenance in production
This example is uniquely connected to the Internet of Things as the manufacturing industry moves in a more automated direction. Perhaps the most prominent example of predictive analytics used in manufacturing is predictive maintenance.
What is predictive maintenance?
The purpose of predictive maintenance is to inform manufacturers of prudent activity regarding industrial equipment. For example, if a conveyor belt in a distribution center breaks down or experiences a malfunction, it can paralyze production and cost the manufacturer money.
By taking large amounts of data, typically through the use of IoT-embedded sensors on the equipment, manufacturers are able to intervene before a crash occurs.
5. Detection of cyber security fraud
More than 3 billion fraud reports were filed in 2018 with the FTC, resulting in $ 1.48 billion in total losses. That’s a 38 percent increase in just one year.
What is one way to tackle the billions of dollars lost due to fraud every year? Well, the use of predictive analytics has become a more prominent solution in the cybersecurity industry.
This is done by analyzing typically fraudulent activity, training predictive models to recognize patterns in this behavior, and finding anomalies. Better monitoring of suspicious financial activity should lead to earlier detection of fraud.
6. Prediction of employee growth in HR
Is it really possible to predict employee success through the use of analytics? The short answer is yes, although HR is still a relatively new industry that takes advantage of predictive analytics.
There are a few ways this can be done. One way is by collecting data to manage workflows and increase productivity. Employee data can show pain points and productivity increases in their day-to-day, and this data only gets better with time.
Using a performance management system to collect this data can help companies predict future employee performance. More data can be used to build baselines for where employees need to be at what stages of their careers.
Predictive analysis can also help during the hiring process. By collecting data on everything from company reviews and social media to growth rates for jobs and development skills, predictive analytics can help recruiters find the right matches for their job postings faster and more efficiently. This can also reduce the turnover rate in the long run.
In fact, application tracking software like Greenhouse is one of the few solutions today that uses predictive analytics and machine learning for just this purpose.
7. Prediction of performance in sports
Professional sports can be fun to watch, but in the end, it’s still an industry where franchises are always looking for ways to gain a competitive edge. The most trendy way to do that now is through predictive analytics.
Baseball has been at the forefront of the use of predictive analytics when it comes to professional sports. It is most common today to predict a player’s future value, along with his regression, based on a complex set of metrics. This helps teams when it’s time to structure expensive contracts.
It’s no wonder why professional sports teams everywhere are on the hunt for data analysts and sports-minded scientists.
Read Wharton’s blog to learn more about how baseball teams in small markets have been able to maximize their budgets using predictive analytics.
