
Campfire 🔥: Using Artificial Intelligence for Marketing
by Emily McMahon Sep 8, 2020

In Conversation with Robbie Adler: Using Artificial Intelligence for Marketing

The Role of Artificial Intelligence in Marketing
First and foremost, we wanted to know how Faraday incorporates AI into their marketing and advertising strategy.
“We use a class of AI known as machine learning,” explains Robbie. “We work with customers to use their existing lead database, and we associate that with a lot of data Faraday has compiled on U.S. consumers. We look at the historical behaviors of your customer and lead base, who they are, what they look like, what they bought from you in the past. We also look at the make-up of who they are and what they do outside of their direct interactions with you as a brand.”
With this information, Faraday trains AI models for specific business outcomes suited to the business’s needs in order to best target advertising campaigns.
It’s a similar form of tech to what you see with Netflix or Amazon and their recommendation engines.
However, this process requires a close examination of the data used to train the models for implicit biases.
“Machine learning models use historical data to make forward-looking predictions. It’s not the algorithm that’s biased, it’s our history,” says Robbie. “When we’re training a model, we make sure we’re not perpetuating things we don’t want to see in the future. That has a lot to do with being very thoughtful about what data you show a model.”
AI involves large-scale pattern recognition, and the reality of our society is that there is a long history of pervasive patterns involving racial, social, and economic injustice. Models can pick up on these, so companies such as Faraday have a responsibility to monitor and adjust the model’s learning to avoid perpetuating negative patterns of behavior.
A Closer Look at the Ethics of Data
One of the most prevalent questions surrounding AI and data is that of ethics. Following recent scandals involving large companies’ misuse of personal data, everyone from individual consumers to government legislative committees are paying closer attention to regulating the access and use of data.
Fundamentally, one component to ethically using data is making sure you have a right to use it. A core component is transparency, and the second is that the people who have your data must take security very seriously.
There are huge amounts of data getting produced every day, and a lot of it actually has limited value as an individual data point. However, transparency and permission gives people an awareness of where their data is going and how it’s being used.
“It’s one of the few areas of bipartisan agreement on the right and left,” Robbie points out. “Generally there’s a sense that big tech is using data in ways that people aren’t aware of, so they’re trying to rein that in. I think Vermont has taken an interesting legislative standpoint to provide some insights into who has consumer data, how it’s being used, and requiring certain levels of security protocols.”
Robbie also acknowledges that with so much data out there, some is bound to be inaccurate, and explains how Faraday mitigates this uncertainty.
“We don’t tell our clients that all the data we have and license is 100% accurate, but we try to find multiple data sources that corroborate one another.”
Following the Evolution of Predictive Data
We wanted to know what Robbie thought about the evolution of predictive data, and where he could see it headed in the future.
If you look at a lot of predictive software, or marketing software that talks about predictions, it really is a rules-based engine that looks at behaviors. A classic example is with marketing automation tools. They’re establishing rules and saying, ‘If this email gets opened, we’re going to set up a certain campaign off of that behavior.’
As predictive data has evolved, however, it has started moving in a different direction for large-scale applications.
“What’s emerging is more of this machine learning-based approach to predictions, which has large-scale pattern recognition and less rules to predict for an outcome. It’s a different approach, and they can often work in conjunction.”
Generating Omnichannel Growth Using Artificial Intelligence
Lastly, we chatted with Robbie about how AI can contribute to omnichannel growth for marketers and advertisers. While platforms like Facebook and Google AdWords allow you to predict an expanded audience based on your existing customer base, that audience can’t be migrated to another platform.
“At Faraday, we’re trying to build a platform that allows you to source an audience that applies these predictions,” says Robbie, whether that’s bringing an audience into Facebook, or even targeting people offline with direct mail. “You’re taking these models that you build around people who are likely customers. Our platform is a bit agnostic as to where we push those audiences, so we push them off of the rules established by those advertising platforms.”
By harnessing these AI capabilities, Robbie and his work with Faraday have helped clients grow an audience far beyond the capabilities of a single, isolated platform. Instead, AI and predictive data provides them insight into potential audiences and untapped consumer bases with the help of the company’s existing data.
In a hyperconnected world where everyone leaves behind what Robbie calls “data exhaust,” understanding where and how our data can be used not only makes us more conscious consumers, but also more targeted marketers.