How Choozle is using machine learning for digital marketing with Callie Federer

Marketing & communications specialist, Sarah Lilly, sat down with Choozle’s data scientist, Callie Federer, to learn more about the possibilities machine learning brings to the digital marketing landscape. During their conversation, Callie shared what machine learning actually is and described some of the exciting new ways it is being used at Choozle.

Machine learning brings a new perspective to digital marketing with a new tool to leverage the massive amounts of data we collect, allowing us to dive deeper into the opportunities hidden within.

Callie received her Ph.D. in Computational Bioscience, where she focused on machine learning in the context of understanding the brain and using the system of the brain to build better machine learning. Since graduate school, she’s focused on applying machine learning in many different fields, including particle physics, cancer treatment, character recognition for a calculus tutoring app, and now helping marketers make better decisions and optimization choices.

So let’s jump in with a few exciting points Callie shared with us about using machine learning for marketing.


What is machine learning?

Machine learning is a way to solve a problem without using a traditional rules-based approach to create software. For those like us, who didn’t know what a rules-based approach is, it’s a way of telling a computer to perform a task in which a data scientist needs to come up with all possible scenarios and tell the computer, if X happens do Y. Machine learning, on the other hand, utilizes data and past experience to understand scenarios with less bias than a human. The example Callie uses is teaching a computer to categorize fruits based on their attributes. A rules-based approach would entail telling the computer, if it’s yellow it’s a banana, and if it’s red it’s an apple. But pretty quickly, you can see this is flawed as there are yellow apples and red bananas in the world. Instead, if you utilized machine learning, you could provide the software with massive amounts of photos and data about fruit and tell the computer to classify the different fruits as best it can. The algorithm can then start to learn things about the relationship between shape and color when it comes to categorizing fruits, so you don’t have to come up with all possible cases yourself.

What’s the difference between machine learning and artificial intelligence?

These two things are talked about and used almost interchangeably. Artificial intelligence is the goal, and machine learning is a possible method of achieving artificial intelligence. From a broad perspective, artificial intelligence is getting a computer to think and act like a human. And while there are some examples of computers performing tasks as well or better than humans, like playing chess, machines still don’t contain the same depth and variety of understanding of things the same way a human does. And to be frank, artificial intelligence doesn’t exist yet. Machine learning is simply a tool we use to help computers learn how to do different things.

What do marketers need to know?

Machine learning is a powerful tool that can help cut down difficult choices and reduce menial work while avoiding inherent human bias. But it’s not magic, it’s just math. While there are a lot of tools that position themselves as automatic machine learning, it’s not always foolproof if there isn’t someone who knows at least some of the math behind the algorithms. The last thing you want is to be making decisions based on an error coming from your machine learning algorithm. You must be able to trust your data and the decisions your algorithm makes for you.

Machine learning at Choozle

The data science group at Choozle is still pretty early in some of the projects we are setting out to achieve, but the first piece of technology we are working on is a budgeting tool for marketing campaigns. This piece of machine learning can’t predict the future yet, but it is helping predict how a marketing campaign will perform based on the previous budget used and past performance of the campaign. You could use this tool to shift your budget up or down and understand how those changes might affect clicks, conversions, impressions, and more. Choozle is helping marketers to make decisions by leveraging years of campaign data to predict outcomes based on the historical trends shown.

We have a lot of other ideas of how machine learning can help you make all of your decisions for your campaigns on the Choozle platform, but as of now, we are still trying to figure out what would be the most helpful to start with when it comes to creating a campaign.

Listen to Callie’s full ChoozleChat podcast here:

If you have any questions or just want to learn more about how Choozle is using machine learning to improve our platform, reach out to us here.

The post How Choozle is using machine learning for digital marketing with Callie Federer appeared first on Choozle: Digital Advertising Made Easy.

Be the first to comment

Leave a Reply

Your email address will not be published.


*