We recently sat down with Kent McCormick Ph.D., co-founder of Lattice Engines and current founder of Photon Vault. Lattice Engines was a first-generation AI/ML platform for sales and marketing, and was acquired by Dun & Bradstreet in 2019.
We talked with Kent about his 15 years at the epicenter of the AI/ML revolution. Check out our interview about successfully operationalizing analytics!
Matt Mesher at Savant: You founded Lattice Engines before AI/ML was even a term. What inspired you to do that and what did you learn from it?
Kent McCormick: What got us interested in AI/ML were the experiences that both my co-founder and I had as consultants at McKinsey and then with tech giants like EMC. We got to see the power of applying analytics techniques to business problems.
We wanted to take those learnings – that others had been applying to areas like supply chain, fraud and risk – and use them in sales and marketing.
We spent a lot of time with our customers discovering how to prepare their data for sales and marketing use-cases, get end-to-end organizational buy-in, and turn it all into something that could transform businesses.
There are many challenges when going from the experimentation stage to an actual, scalable solution that your business can depend on. Some of them are technology-based and others are organizational in nature. Both have come very far in the last several years.
Matt: What does successful analytics look like in your opinion? What do you look for?
Kent: The critical thing is to make sure you're focused on an impactful business problem and understand how the analytic insights will be used in business operations.
In my experience, frequent and fine-grained decisions make the best opportunities for analytics impact on the business.
Matt: What were some of the challenges you see teams facing when trying to operationalize analytics?
Kent: One of the biggest is just getting the organization to accept the analytics and start running the business in that way. People are already used to working one way and you need to create buy-in for your analytics.
If you don’t have the commitment top down and bottom up in the organization to actually embrace the analytics, then it can be hard to have a successful project. You may build something cool, but that doesn’t mean it will get used.
The other thing, which is always an issue, is the data.
There have been situations where there’s a lot of excitement and interest from everyone in the organization, but when it comes time to actually pull the data together, there’s no cross-organizational data infrastructure. So what you imagine is easy, actually becomes very hard. It certainly can extend the timeline and costs, and, in some cases, make it impossible to execute the vision.
Matt: What is one piece of advice you have for someone looking to build a highly impactful data and analytics org?
Kent: Do not underestimate the challenge of understandable, trustable and maintainable analytics.
This is ultimately about how you make sure what you’re building does what it’s expected to do and keeps doing what it's expected to do and allows for incremental improvements (without breaking ten other things). Without infrastructure that’s designed for this, you end up going down a spaghetti path that limits your capabilities over time.
The problem is that when your teams are focusing on just maintaining various parts of your data stack, then your analytics suffer. No-code analytics automation platforms like Savant can be hugely helpful for this – simplify the experience and focus on the analytics.
This is a golden age for analytics, especially with the advent of these new tools and technologies. It’s a very exciting time to be an analyst and certainly the most exciting time in my career.