Machine Learning for the Rest of Us

By | 2016-11-09T15:10:16+00:00 November 9th, 2016|Azure, Azure Machine Learning (AML), Data Science, Reporting, SQL Server|1 Comment

Machine learning. Everyone is talking about it, but most have no idea what it is or why it is needed. Machine learning may seem like magic, but if one does not know where to start they can quickly get lost. All organizations have a need for predictive analytics because it allows for quicker and better decision making. Machine learning is at the heart of modern business and can give you an edge in an ever-changing market.

All companies collect data. Data about sales, customers, employees, and competitors. The ability to utilize this data by making quick and automated decisions can radically elevate an organization’s performance. Using available data, machine learning employs mathematical algorithms to model and predict outcomes. Companies can gain fast actionable insights by predicting values, classifying customers, recommending services, and detecting abnormalities. These machine learning solutions enable organizations of all sizes to work smarter, faster, and more efficient than ever before. Here are a few examples of machine learning models which may help an organization improve:


Regression analysis processes different characteristics of an event, customer, or product to produce a numeric value. For example, if a company would like to estimate the return on investment for a specific marketing effort, then they could use a regression model to predict an approximate dollar amount, representing the return for the marketing strategy. Being able to accurately predict revenue will let an organization compare the actual revenue with the expected revenue. This allows for further evaluation of how or why the marketing strategy fell short, met, or exceeded expectations.


Properly classifying new customers, products, and events into predefined categories can save time, money, and energy. Take a bank for example, a loan officer wants to process a customer’s information and predict if they are a credit risk. A classification model would be built using previous customer information, learning the characteristics of individuals who are high risk and individuals who are low risk. This model enables loan officers to process customer information and more accurately predict which customers are risks, and the probability of that customer faulting on their loan.


When one thinks of recommendation engines they often picture online retailers displaying ads for other products the customer may be interested in. This is a great example of a recommendation engine that automates the suggestion process, bringing in extra revenue for the company. However, recommendation engines not only recommend products, they can also be used to generate work flow for employees. For example, suggesting the next best action an employee should take is a great way to proactively interact with customers. Should an employee send an email asking for a review of services, or should they call to pitch another product? Product recommendation and “next best action” are just two examples of how a recommendation engine can help organizations get to the next level.

Anomaly Detection 

Anomaly detection is the identification of events or products that do not fall in line with the expected pattern of the data. The most popular example of this is fraud detection. Banks and financial institutions across the globe apply these machine learning algorithms to monitor the purchases of their customers. When a purchase does not fall in line with the customer’s typically purchasing habits there is an alert questioning the transaction, often in the form of a call, text, or email. Anomaly detection can also be used in monitoring equipment. Sensors can be installed on hardware to monitor the condition of machines, and when there is an unexpected performance change a notification is sent to service the equipment before it breaks entirely. Anomaly detection allows for recognition of rare events enabling fast response when problems arise, saving time, energy, and money.

Being able to gain valuable insights from data will allow a company to grow and improve. Foresight is critical in the modern marketplace, and employing machine learning is the best way to obtain it!

Machine Learning is a key part of modern marketing organizations.  So is good data. Northwest Cadence has an offering to fit all sorts of needs.  If you are curious about how you can use machine learning to transform the way you work with your data and do business, please reach out to and we’ll work with you to create just the right solution for your team!

About the Author:

One Comment

  1. Steven Borg November 11, 2016 at 8:38 am

    This is a really good summary of the various techniques used in data science. Thanks for pulling together the non-obvious example uses of each. For instance, “next best action” as a recommendation engine. That’s a great idea!

Leave A Comment