Session library

AI-Powered Marketing Insights

Judah Phillips · July 5, 2024

ai powered marketingattribution
**(04:24)** Judah: Sure, happy to share. A quick background: I’ve been in the internet and software space since the mid-90s. My first startup was around 1997, working in information retrieval. It was before Google, back when Northern Light was big. I worked with indexing data from the Congressional Record, building search tools and web interfaces.

Later, I moved to Boston, where I’ve been since, and worked for companies like Sun Microsystems and Monster Worldwide. Over the years, I realized that consulting and service businesses, while rewarding, are hard to scale. I wanted to focus on software, so I co-founded Visad Data with my partner, Dan. We initially built data science tools focused on machine learning, like clustering and market basket analysis, mainly for businesses.

Our big product became **Squark**—a machine learning tool that could do binary classification, multinomial classification, regression, and time series prediction. We named it “Squark” because it sounded catchy, and the word itself is the super-symmetric partner to the quark in physics. We made rookie mistakes starting out, but ultimately grew it into a solid product that enterprises could use.

**(10:36)** After some initial success, we started selling Squark to companies like IBM Watson, Nealon, and Epic Games. We worked on major projects, like churn prediction for Fortnite and liver donation campaigns for healthcare organizations. We competed with major players like DataRobot and Microsoft and managed to succeed through a lot of innovation and bootstrapping. 

Eventually, a 4,000-person, 90-year-old company operating across five continents acquired us. On May 3, we completed the acquisition, and now I’m their Chief AI Officer. My goal is to bring responsible AI into the company, focusing on four areas: team development, software solutions, data governance, and infrastructure.

**(17:03)** I live in Boston with my family, enjoy concerts, and try to ride my bike when I can.

David: So, as Chief AI Officer, what does your team look like?

Judah: We’re still building it out. Key roles include full-stack, front-end, and back-end engineers. DevOps is part of engineering but may include network operations for scaling and security. We have analysts who use or code our technology, data engineers for managing data in Snowflake, and sales staff to help bring our solutions to clients. We don’t just focus on algorithms; we’re also hands-on with clients, helping them leverage AI for predictive insights and decision-making.

**(22:01)** David: How do you interact with the marketing team? Are they also interested in applying AI?

Judah: Marketers vary in their readiness. Some are open to AI and advanced analytics, while others are more hesitant. Over time, I learned to focus on the marketers who are ready to learn and invest. I ask straightforward questions to understand their buying process, as that determines how we proceed. We offer a 90-day trial for those genuinely interested, where we onboard them and train them on the platform to ensure they can generate value by the end.

One challenge we faced was competing with major players like DataRobot. We worked with big clients on predictive models, but without the huge support teams that larger firms have. However, we built effective software that often outperformed larger competitors’ models, proving that a smaller, focused team can succeed in enterprise spaces.

**(28:27)** David: How do you decide which AI tools or technologies to incorporate?

Judah: We were a small team, but we focused on listening to our customers and responding to their needs. We built Squark based on customer feedback, creating features they requested and enhancing the platform with their input. By being more agile and responsive than larger companies, we could provide real value that was tailored to each client.

Building customer trust was also key. Selling the company taught me that honor and integrity are invaluable. You want to work with people you can trust, and that’s true for clients and investors alike.

**(32:06)** David: Looking at the future of search, do you think we’re facing a major shift, especially in terms of Google’s role?

Judah: I avoid AdWords myself, but I see the value in Google’s platform. With tools like ChatGPT, search habits may change, but I wouldn’t bet against Google. They have the resources and the drive to adapt, and I think they’ll continue to dominate. It’s an oligopoly—hard to break into, even with VC funding. Still, new players like Perplexity are emerging, and they could create niche opportunities or become acquisition targets. 

**(40:13)** David: My concern is more about companies that rely on Google for customer acquisition. A 25% drop in search could disrupt a lot of businesses.

Judah: Absolutely, and I get it. Some industries, like recipe sites, are feeling the squeeze as people use AI for customized results. While I’m personally okay with skipping those ad-heavy sites, I understand it impacts the people behind them. It’s going to be interesting to see how everyone adapts if search models continue to shift.

**(43:32)** David: You’ve mentioned Squark’s use in nonprofit spaces. Is that still a focus?

Judah: Yes. We’re working with nonprofits to improve fundraising strategies. Squark enables predictions about donation behaviors—who will donate, when, and how much. For example, we’ve helped clients like the University of Pittsburgh Medical Center use Squark to predict which individuals are likely to donate organs based on lifestyle and other data.

**(47:27)** David: In terms of big wins, where do you see the most opportunity?

Judah: Three main areas come to mind: customer journey mapping, media mix optimization, and attribution. The customer journey involves predicting customer behavior at each step—conversion, retention, and loyalty. Media mix optimization is another powerful area, where we analyze multiple ad channels to predict ROI and optimize spend. And finally, there’s attribution, where AI helps pinpoint which data attributes most influence conversions.