Predictive Analytics on the Rise: A Q&A with MediaAlpha Co-Founder and CEO Steve Yi

Predictive analytics is one of the biggest trends in digital marketing.

The rise of predictive analytics is one of the biggest stories in the marketing technology world, with brands across a wide range of industries building predictive models to forecast future outcomes based on their historical data. In a recent interview with MarTech Series, our co-founder and CEO Steve Yi discussed this key trend, as well as several others that are shaping the way businesses connect with their target consumers online. The interview is re-printed here.

MarTech Series: Tell us more about MediaAlpha and how the platform has evolved the last few years?

Steve Yi: We are the leading customer acquisition platform for the insurance industry. Our technology solutions give our partners the ability to identify and acquire the right customers at the right price. As to how we’ve evolved the platform, over the years we’ve introduced insights, tools, and controls that provide full transparency into where consumers are shopping online for insurance, as well as all of the data that consumers provide during that shopping experience. This enables our partners to more precisely target the consumers they wish to attract, and to ensure that the price they’re willing to pay allows them to achieve their business goals. We’ve also incorporated lead-edge technologies such as predictive analytics and AI into our platform, to help make data-driven decision making as informed and efficient as possible.

MS: We’d love to hear your thoughts on the current top trends surrounding the evolving predictive analytics landscape in martech and a few top providers from around the world that you’d like to highlight, those that are changing the game?

SY: Predictive analytics boils down to using historical data to make better future decisions. You build a model using your historical data, you learn from that model, and you refine your strategies going forward. So it doesn’t have to be thought of as a product you buy from a provider. Brands can do this themselves and bake it into their technology stack in various ways. In our case, predictive analytics is a core capability we offer, and it’s intertwined throughout our platform. We work very closely with our partners to model their data (if they don’t have that capability in-house) and apply the insights into their customer acquisition and monetization strategies. As brands become more and more sophisticated as to the consumer data points that matter, i.e. the ones that actually signal purchase intent, predictive analytics will help them become exponentially more efficient in acquiring new customers.

MS: Specifically, how are you seeing predictive analytics change the game for insurance marketing today?

SY: When people think about predictive analytics in a marketing context, they typically think of the way it can be used to help companies identify their most likely customers. But one thing we’ve found in insurance is that carriers can actually benefit a great deal by using this technology to identify their least likely customers. For instance, let’s say a shopper arrives at an insurance carrier’s website and fills out a form to receive a quote. Based on the data that consumer provided, the carrier’s predictive model suggests the consumer is highly unlikely to purchase a policy. That’s a shopper that the carrier worked hard—and spent significant marketing dollars—to bring to their website, only to have them leave without purchasing a policy. But if the carrier shows them offers from alternative carriers alongside their own quote, they can monetize this consumer.

Predictive analytics is the key here because it allows carriers to make smart decisions about which shoppers to show these alternative offers to. After all, if a shopper is highly likely to purchase a policy, you wouldn’t want to distract them with an offer from another carrier. But if they’re highly unlikely to purchase a policy, you don’t have to worry about sending them somewhere else, since your predictive model has made you fairly confident that they weren’t going to purchase a policy from you anyway.

This strategy has helped some carriers earn back as much as 30% of their overall digital marketing costs. That’s potentially millions of dollars of new revenue that they can retain to increase the bottom line, or invest in customer acquisition to bid more aggressively for their target consumers.

MS: What best practices can you share with marketing teams when it comes to optimizing their data and predictive insights to identify and target high value customers better?

SY: It all starts with having enough of the right data. In order to make the most of predictive analytics, carriers need about two to three years of historical customer data. This makes sound data collection practices a must.

The other thing we’ve found to be useful is bringing all of the relevant stakeholders to the table, so that everyone on your team can have confidence in the predictive model you’re building. After all, you can have quality data and a great model, but if the people on your team don’t trust the model enough to make decisions based on its predictions, it’s not going to do you any good.

MS: How do you feel predictive trends will evolve for all marketers across industries in the near-future?

SY: As the world continues to shift to digital, two things are happening: brands are getting access to more consumer data and marketers are becoming more sophisticated in how they analyze and apply that data to make smarter decisions. In short, it’s only going to become more feasible for marketers to employ predictive analytics.

As that happens, I think we’re going to start seeing more brands bringing their programmatic media-buying operations in-house. Because a brand’s consumer data is so central to predictive analytics and other forms of data-driven decision-making, it just makes a lot more sense to have the people who are analyzing the data under the same roof as the data itself.

I think agencies deliver a lot of value on the creative side, but marketers will increasingly realize that they’re better off without the friction of sending their data back and forth to an outside media-buying partner. As new privacy laws place greater restrictions on what brands can and can’t do with consumer data, the benefits of in-housing this work will only become more apparent.

MS: A few predictions that you have for the future of martech and adtech?

SY: All signs point to the growing importance of contextual and first-party data. With a growing regulatory focus on privacy around the world and Google’s forthcoming removal of cookie support from Chrome, marketers have no choice but to look to other kinds of information to identify and price their target consumers. In this environment, brands will need deep transparency into the context of every click if they want to maximize their performance.

MS: A few takeaways for marketing leaders and CMOs/CEOs in 2021: top factors they should keep in mind as they plan for the rest of the year, innovate and expand their teams?

SY: One trend we’ve really been noticing is a change in the educational backgrounds of the people who work in programmatic marketing. We’re seeing the rise of a new breed of data-driven marketer that you just didn’t encounter much even five years ago. These quantitative, highly analytical practitioners are coming from the worlds of data science and finance, and many have advanced degrees in those subjects. That’s in stark contrast to the more traditional profile of people who came from marketing or broad, liberal arts backgrounds.