Welcome back to The Forecast Files, everyone. I’m Claire Whitmore, and as always, I’m joined by the ever-insightful Henry Sinclair. Today, we’re diving into something that’s become a bit of a buzzword in our world—PyMC models. But, you know, not just for the data scientists in the back room. We’re talking about digital intuition for everyone.Absolutely, Claire. And I love that phrase—digital intuition. It’s not about replacing gut feeling, but rather, giving it a mathematical backbone. With PyMC, we’re not just saying, “This campaign will work,” but rather, “Given what we know, here’s how likely it is to work, and here’s how confident we are.” It’s a subtle but powerful shift, isn’t it?It really is. And it’s funny, because when I first encountered PyMC, I’ll admit, I thought it was just another tool for the coders. But then I realised, it’s actually about making statistical thinking accessible. I remember sitting in a meeting—this was years ago now—and suddenly, the conversation wasn’t just about numbers, but about what those numbers meant for our creative and strategy teams. It was a bit of a lightbulb moment for me.That’s brilliant. And it’s a big departure from the old way, isn’t it? Traditional forecasting—well, it’s often a single number, a point estimate. “We expect 10,000 clicks.” But with PyMC, we’re saying, “We expect somewhere between 8,000 and 12,000, and here’s how sure we are.” It’s a range, a probability, and it’s so much more honest about uncertainty.Exactly. And as we discussed in our last episode, embracing that uncertainty actually builds trust—with clients, with teams, with everyone involved. It’s not magic, it’s just math, but PyMC makes it feel approachable. And that’s what I love about it.So, let’s get into how these models actually come together. I was listening to Priya from the data team the other day—she put it so well. You start by asking, “What are we trying to predict?” Then, “What data do we have?” and finally, “What assumptions are we making?” It’s almost like building a story, isn’t it?It is. And you don’t need to be a coder to follow the logic. I remember a case where a marketing team wanted to forecast the outcome of a creative campaign. They didn’t write a single line of code. Instead, they mapped out the variables—budget, audience, timing—and the relationships between them. PyMC let them express all that, and then simulate thousands of possible outcomes. It was simple logic, but incredibly powerful.That’s the beauty of it. When you understand the logic behind the model, you can start to challenge assumptions. Maybe someone says, “Oh, we always get a 5% click-through rate.” But with PyMC, you can test that. What if it’s lower? What if it’s higher? Suddenly, the whole team is collaborating, asking better questions, and making smarter decisions.And it’s not just about the numbers. It’s about the relationships between them. If you can imagine a scenario, you can help shape the model. That’s what makes it so inclusive. You don’t have to be technical—you just have to be curious.Right, and as we’ve said before, the more people who understand the model, the better the outcome. It’s not just a data team thing. It’s a whole company thing.And that brings us to something I’m really passionate about—empowering non-coders. PyMC models aren’t just for the technical experts. They’re for thinkers, planners, creators—anyone who wants to make smarter decisions. I’ve seen cross-departmental teams use PyMC to simulate thousands of campaign scenarios. Suddenly, strategy, creative, and client services are all speaking the same language.I’ve got to share this—there was a workshop I ran not long ago, and honestly, I thought the non-coders would be a bit lost. But they surprised everyone. They built insightful models, asked the best questions, and really challenged the data team. It just goes to show, if you understand the goal and the logic, you’re already halfway there.Absolutely. And I think that’s the real takeaway. If you can imagine a scenario, you can help shape a model. PyMC is about making data-driven decision-making accessible to everyone, not just the people who write code.Well said, Claire. And I suppose that’s a good place to wrap up for today. Next time, we’ll dig into a real-world case study—how PyMC helped a team optimise their media strategy. Should be a good one.Looking forward to it, Henry. Thanks for joining us, everyone. Stay curious, and we’ll see you next time on The Forecast Files.Goodbye, Claire. Goodbye, everyone.

Episodes

Ep.3 - Digital Intuition with PyMC