Machine Learning Basics for Mobile Marketing Teams

Updated on November 16, 2018
Machine Learning Basics for Mobile Marketing Teams

Ever wonder how Gmail or even Outlook knows when an email is spam? How does it automatically categorize messages as worth trashing?

Or how does Amazon know to recommend buying socks and foot powder to someone browsing shoes?

The answer in both cases is machine learning. And it is the key to growing your app and marketing it at scale.

In this basic introduction to machine learning, we cover examples, benefits, and how machine learning can lead to marketing automation.

Introduction to Machine Learning

Machine learning (ML) is the use of artificial intelligence (AI) to give a computer system the ability to learn by itself. The process makes machines more intelligent over time. And it allows computers to adapt to human behavior.

The computer system learns from past data that we feed it. Some of that may occur at the start, with a person pointing out which data is useful. Some of it may happen when people give the system data points that help it make decisions later on.

The outcome: the computer system can make predictions and decisions on its own thanks to the data.

Machine Learning Examples in Everyday Life

You may have been using machine learning already without realizing it.

One of the best machine learning examples is Facebook’s facial recognition feature. Tag a friend in a photo once and you may never need to tag her again. And if the system isn’t sure, it asks. Facebook learns from every person tagged which name fits which face.

Then there is Netflix’s recommendation engine. The more you watch and rate movies, the more data it has to recommend movies that you might also like. Every time you rate a movie, you feed data to Netflix’s system. Every time you start a movie or show then abandon it, you provide another data point. The result: Netflix knows what you do or don’t like watching, and what you enjoy bingeing! Thanks to machine learning.

Paypal has a feature that allows it to detect whether a transaction is real or fraudulent. Their system can identify when you are using the account and when it looks like someone else has used it. Yet another of the many machine learning examples we rely on every day.

Machine learning basics - ML can help segment your millions of users

Machine Learning Basics: The Different Types of ML

There are three major types of machine learning:

  • Supervised learning is about making predictions. This is where the system, after enough training, has the ability to predict outcomes. Even when it comes across new data, it can work the problem out. It will incorporate the new data into the system using the rules it started with.
  • Unsupervised learning is about finding patterns. Here, the system uses algorithms when it meets new data. It has to draw inferences and find hidden patterns, similarities, or anomalies in the data. Then it can find relationships between this new data and existing data.
  • Reinforcement learning is about iterative learning based on past results. Here the system learns as it cycles through decisions. Is the decision good or bad? It can use the result of a current decision to inform the next one. We use this type of machine learning to increase the efficiency of a computer program.