We sat down with Ron Szymanski, Senior Director, Data Science & Analytics, to better understand how he translates big data and analytics into meaningful insights for our clients.
How do you make analytics accessible to people who aren’t “numbers” people?
For me, it all starts with speaking the same language and asking the questions that really matter. How does your business make money? What are your goals and where do they fit inside the organization? By understanding these things, we understand what matters most – new accounts, new patients, new partners, etc. – and can focus on connecting those things to actual goals. For example, if we add X amount of new customers what does that look like for revenue. By continuing to connect the dots, we can figure out that revenue means Y amount of profit based on this margin.
At the end of the day, no one wants to listen to me drone on about media platform key performance indicators, compared to website performance indicators, and how those compared to a forecast or model that was built based on a set of performance assumptions. The end game is how many of X did we bring in and how much Y did it generate. There may be a million data points in between, but those two numbers are the ones that matter most and tell us what our work is doing for the business.
How do you help clients separate signal from noise when they’re drowning in data?
The sheer volume and number of different data sources make the noise topic way too real. I like to bucket or segment data by its purpose, where what is being done lands in the marketing funnel, and how it is measured. Is this awareness data? If so, we measure it via frequency, reach, engagements, and studies. Is this consideration data? In that case, we measure it by interactions on the website from different media platforms, or organic/direct visitors to gauge intent and if we are in the consideration set. Is this conversion data? Then how many leads, forms, or accounts were submitted and where did they come from and what happened after the sale or lead. There is a lot of data, but in the end it comes down to where the data is coming from and having a clean framework of how to interpret the data.
What’s the biggest analytics mistake you see marketers making right now?
The biggest mistake I see is that marketers often treat analytics as a reporting function rather than a growth engine. If you only have your analysts in rear-view mirror mode, then you’re missing out on the huge opportunities to proactively shape the results. Through experimentation and testing, we learn as we go and can we use modeling to predict how these experiments will move the needle. We can engineer better outcomes by making changes and testing along the way. Learning a piece of creative or webpage is not working should not be a post-campaign item; it should be challenged constantly for continued success.
How is AI changing the analytics landscape? What does that mean for analysts in 2026?
Let’s face it, AI is transforming everything we do as analysts. The rise of ChatGPT, Perplexity, Gemini, Claude, and more has made the large language model available to anyone who can ask a question and upload a spreadsheet. AI has eliminated a lot of grunt work and makes it an exciting time to be an analyst. The key is to find the sweet spot between not exposing client data directly to LLM’s and speeding up work and allowing more complex analysis than would have been possible if you were writing code line by line.
As far as 2026 and beyond, I hope one day we can trust the math skills of the AI models. Currently, you can’t. There are errors in simple calculations and often, with missing business context, AI gives the wrong analysis of what’s positive or negative. However, every time a user or analyst points out mistakes, the models are learning and eventually they will get better at confirming complex calculations.
AI won’t replace analysts in 2026, but those who use AI will replace those that don’t.