What is Machine Learning?

AUGUST 8, 2017

Machine Learning is an application of Artificial Intelligence (AI) that takes large amounts of data and searches for patterns and insights that improve decision making. What calculators did for math, machine learning does for data. The primary goal is for computers to learn without human interference, adjusting and maturing as new data is introduced.

“What calculators did for math, machine learning does for data.”

Say you’re a retail company, and you want to know whether you should send a customer a coupon. You have a rich database full of historical customer data that includes demographics, spending history, and whether or not the person redeemed the coupon. A machine learning program will go into this data set and look for what affects the output, specifically, “Will the customer use the coupon?” The algorithm has the ability to uncover unrecognized key performance indicators (KPIs), relational trends between fields, and predictive formulas to find the optimal customer to target. As more data is added, the system takes its prior knowledge and adjusts its predictions based on the new information.

This type of machine learning is known as “supervised” because the program is given the directive, “Will the customer use a coupon?” Machine learning can also be “unsupervised,” which means the data is unlabeled and the AI is not told how or where to search. It is simply turned loose and asked to uncover hidden connections between metrics as well as trends within the data.

Take another grocery store example: There are three coupons for beer, diapers, and lunchables. An unsupervised program could look at who was most likely to use each of these and then create “clusters”. Those clusters could be predictable—mothers with three children are most likely to use the lunchable coupon—or they could be a completely new insight, like dads who go pick up diapers also grab beer. As a result, the grocery store can target dads with coupons for diapers, knowing that they will also purchase beer.

It’s important to understand machine learning is founded in historical information. Thus, it adds value to companies by producing inferences and insights that would take humans much, much longer to discover themselves—if they even have the capability to detect them.


CoEFFICIENT® utilizes machine learning by taking historical data from across the enterprise and surfaces relevant information throughout the platform. These insights help track business health, identify actionable opportunities and empower users to take action. The platform’s background data is analyzed for trends, gathering both historical and cyclical patterns. The machine learning algorithms then establish thresholds for future data so that outliers are discovered whenever data dips below or rises above the historical marker.

Recommended Metrics are the most flexible CoEFFICIENT output and also one of the most powerful. They leverage machine learning to gather usage patterns over time and serve up a personalized suite of metrics for each user in the company, depending on the person’s role and needs. Recommended Metrics will also display rarely checked KPIs whenever they are approaching or exceeding a predetermined threshold. In essence, the machine is telling the user, “Attention! This is abnormal, so you may want to look into it.”

Related Metrics utilize machine learning in a correlative way. While still supervised, it applies aspects of unsupervised methodology. Upon clicking into a metric, the user will find several more metrics that are connected to the KPI. These are uncovered through machine learning algorithms that find and explore connections between different business practices. A District Manager, for example, can click on a metric that tracks NPS and see how a store benchmarks to industry or company standards in checkout times, associates on the floor, and other practices that may affect customer experience.

Recommended and Related Metrics are just two out of the many ways that the CoEFFICIENT platform utilizes machine learning. The power of this artificial intelligence is built in from top to bottom. By tapping CoEFFICENT to do this work, District Managers are freed up to focus on the more valuable aspect of their role: coaching store managers on performance and execution. Read our Metrics Data Sheet to see how CoEFFICIENT can help your company understand it’s KPIs.


As a member of our marketing team, Zach helps shape up the content calendar and keeps operations running smoothly. He loves telling stories that entertain, sharing information that engages, and creating content that empowers. Ready to tackle anything that comes his way, Zach’s motto is always, “Absolutely. I’ll give it a shot.” Outside of Square Root, Zach loves to travel and has seen 43 of the 48 contiguous states. An avid basketball player, he describes his jump shot as “picture perfect.” Zach will graduate from the University of Texas at Austin with dual bachelor’s degrees in marketing and radio-television-film in December 2017.