Home blue house icon orange house icon
Home
Contact Us blue email icon orange email icon
Contact Us
Search blue search icon orange search icon
Search
IDENTITY
REVOLUTION
<< View All Episodes

Episode 2 –
Using AI to Predict Consumer Behavior in an Ethical Way With Andy Rossmeissl of Faraday

RESOURCE CENTER   ❯   Identity Revolution Podcast 12-16-20

Episode Summary

It didn’t take long for Andy Rossmeissl to realize AI’s potential for predicting most consumer decisions.

As the co-founder and CEO of Faraday, an end-to-end AI platform that transforms customer data into actionable predictive insights, Andy’s AI software has “made hundreds of billions of predictions for hundreds of consumer brands.”

“AI is really just the newest and best way to solve problems that growing companies have faced for decades,” Andy says. With huge companies like McDonald’s already spending $300 million on AI to help customize their menu for each customer, brands can’t ignore prediction software for much longer.

Using software to predict consumer behavior is becoming a necessity for modern-day consumer businesses, but there’s a right way and a wrong way to approach AI.

On this episode of Infutor Data Solutions’ Identity Revolution, Andy talks about how he founded a data company built on ethics (yes, it’s actually possible!), what e-commerce brands can do to stay connected with customers, and how businesses can extract the most valuable data possible.


Guest-at-a-Glance

  • Name: Andy Rossmeissl
  • What he does: As the co-founder and CEO of Faraday, an end-to-end AI platform that transforms customer data into actionable predictive insights, Andy helps hundreds of consumer brands create an unforgettable customer experience through the power of AI.
  • Company: Faraday
  • Key Quote: “Once we recognize that privacy is an asset that every individual has, and can feel free to monetize however they want, we’ll end up with a much stronger, more sustainable martech and adtech ecosystem.”
  • Where to find Andy: Twitter | LinkedIn | Personal Website

Key Insights

  • Customer (and brand) identity is crucial. Andy sees a rapidly increasing danger with the rise of e-commerce, especially during COVID, where businesses no longer see their customers as individuals. Without the face-to-face interaction of a customer walking into your store, making small talk, and building a positive brand experience, it’s easy to lose sight of who your customers are and what their largest needs are. Yes, tracking customers online with cookies is one way to figure out who they are, but the best way is to participate in the economy — consumers are more upfront about their needs and wants than you might think.
  • First-party data is more valuable when combined with customer profile data. While first-party data and customer profile data hold some value on their own, Andy believes they’re more valuable when combined: “together, it’s like a one plus one equals three type of arrangement.” Analyzing both datasets in tandem makes it easier to preemptively predict a customer’s actions.
  • Don’t make assumptions based on non-existent data. In digital marketing, it’s easy to create superstitions about potential customers. For example, if you’re going after a target demographic your brand has never pursued before, they won’t be in your data or show up in existing buyer personas. When this happens, Andy recommends creating a reliable experiment, backed by a “falsifiable hypothesis” and, of course, strong data so you can run a test on said new business venture or market segment.

Episode Highlights

AI is the newest and best way to solve decades-old problems

“The short answer is that we predict human behavior — like I said, we make billions of predictions every day for hundreds of brands. The longer answer is that AI is really just the newest and best way to solve problems that growing companies have faced for decades or longer. Things like knowing when to discount and who to discount, or treating each customer like an individual, finding your next buyer. These are not new challenges, it’s just that — now at 2020 going into 2021 — it’s fairly clear to practitioners in this space that prediction is the correct way to solve all of these problems at this point in time.”

On creating a company based on unwavering ethics

“We’re really founded on ethics here. There is a right way and a wrong way to do AI. We have technical safeguards against algorithmic prejudice. We always have — and always will refuse — to ever scrape data off the web, which was a very odd decision back in 2012, but it’s one that we made for ethical reasons that we stick to. We also have a policy here against never using cookies or device IDs.”

Addressing the challenges of data hygiene

“Prediction is the ROI story for big data — it’s the reason why we’ve been doing this for so long … and obviously the better hygiene you have, the better your predictions will be. So in that sense, I would say that we emphasize the need for data hygiene and, in fact, we make data hygiene valuable. It’s always been one of those things that is at the bottom of everybody’s list … and that’s fair, because hygiene for hygiene’s sake isn’t really that valuable. However, if you know that you’re taking your data from under your mattress where you’ve kept it for years and you’re putting it in an interest accruing account, which is what we like to think of predictions of providing, then suddenly it becomes really valuable.”

Finding the right approach for collecting private data

“I think society as a whole is really starting to grapple with the extent to which the data that they create, in private, is being used for some good reasons and other not so good reasons, but there needs to be a bigger conversation about that. We’re not saying we can’t do it, but I think we have to be a little bit more careful about how we do. So when you look at the martech and adtech stack, what percentage of those companies really rely on third-party tracking cookies? I think it’s a massive percentage. So I think that many of these folks will find the right way to do this and we’ll pivot and we’ll get there.”

Set aside assumptions about your customers and focus on the data

“The number one thing I would do as a CMO these days is just making sure that, every single day, I know who my customer is. There is a tendency, especially in the digital world, to develop some very strong superstitions about your customer base and not just develop them, but also act on them and, in doing so, also entrench them. Your superstitions can become reality if you decide that your customer is one person and you spend a bunch of money against that one person, you’re going to self-fulfill your prophecy. Maybe that’s the right move, but maybe it’s not. You’ve got to gather data and put your superstitions aside for a second.”

To scrape or not to scrape?

“Everybody is very upfront about their needs and wants as a consumer and that data does get ethically captured by organizations like yours and permission is sought. There’s a very long history of participation in the marketplace, serving both the buyer and the seller. We started the company in 2012 we were like, ‘well, should we scrape or should we license?’ and I think 99% of tech startups back then would have — and in fact did — choose to scrape data. We feel that, working with companies like Infutor, we’ve been able to avoid a lot of those pitfalls and I think position ourselves much better because of it.”

Brands care about the results of AI, not the technical intricacies

“We include everything to make predictions. So every other piece of AI software out there is all about building pipelines, building models, feature engineering, cross validation, all kinds of arcane details. But not every company out there can have a huge team of data scientists operate these platforms. We believe that brands care about the results of AI, not so much how it’s made.”


Top quotes:

[3:20] “We realized that the whole process of predicting consumer behavior and using that information to get closer to your customer and grow more effectively is a pretty widely applicable need out there for consumer brands. Since then, we’ve made hundreds of billions of predictions for hundreds of consumer brands.”

[6:14] “The world is changing and I think it’s changing in this direction. It’s not easy to make predictions ethically, and it’s something we have to keep our eye on.”

[16:47] “I think society as a whole is really starting to grapple with the extent to which the data that they create, in private, is being used for some good reasons and other not so good reasons, but there needs to be a bigger conversation about that.”

Want to speak with the identity data experts?

Contact Us