The Analytics Power Hour

If you've ever poured months into building a semantic layer only to watch it become shelfware the moment the business pivoted, Jacob Matson has some thoughts. And a metaphor. Your data is a jungle—and a semantic layer is a highway. Great if you need to get somewhere fast and reliably (monthly active users: highway, please). But the interesting business questions? The slicing, the dicing, the nuanced dimensions that actually differentiate your company from its competitors? There's no highway for that. There never will be. Jacob, a developer advocate at MotherDuck with deep roots in accounting and ERP systems, joined Michael, Moe, and Julie to talk through what comes after the semantic layer—or at least alongside it. The conversation covered why the most important parts of any business are precisely the parts that resist being modeled in someone else's framework, why AI is actually pretty good at writing SQL but not so great at remembering what it figured out yesterday, and whether the real job to be done here is less about modeling and more about search. Oh, and the uncomfortable truth that at episode 300, we still don't have a great answer for metric drift. But we've got some really good questions. For complete show notes, including links to items mentioned in this episode and a transcript of the show, visit the show page.

Picture this: four analytics professionals, one live audience, a bunch of submitted questions, and absolutely no filter when it comes to sharing their real thoughts about AI, stakeholder management, and the state of the industry. That's what you get when the Analytics Power Hour goes live from Marketing Analytics Summit, with Michael, Moe, Tim, and Val fielding everything from, "How do I prove I'm a partner rather than just an order taker?" to "What's your icky threshold with AI?" The conversation ping-ponged from the fundamentals—like why curiosity beats feature checklists when selecting tools—to the controversial, including a heated debate about whether AI-generated meeting notes are helpful productivity boosters or lazy crutches that strip away human editorial judgment. Along the way, they tackled data trust issues, the pressure to show AI efficiency gains, and why trying to nail down the "best" deliverable will just trigger existential musings about what a deliverable even IS! Fair warning: Tim gets triggered by AI hype, Moe calls some industry BS, and everyone agrees that being useful beats being right. For complete show notes, including links to items mentioned in this episode and a transcript of the show, visit the show page.

Semantic layers are having something of a moment, but they're not actually new as a concept. Ever since the first database table was designed with cryptic field names that no business user could possibly understand, there's been a need for some form of mapping and translation. Should every company be considering employing a semantic layer? Is the idea of a single, comprehensive semantic layer within an organization a monolithic concept that is doomed to fail? These questions and more get bandied about on this episode, where we were joined by industry legend Cindi Howson, Chief Data & AI Strategy Officer at Thoughtspot. For complete show notes, including links to items mentioned in this episode and a transcript of the show, visit the show page. This episode's Measurement Bite from show sponsor Recast is an explanation of multicollinearity from Michael Kaminsky!

As the world turns, a couple of things happen: 1) we grow and learn, and 2) the world changes. On this episode, inspired by a job interview question, the hosts walked through a range of thoughts and beliefs they had at one time that they no longer have today. Analytics intake forms are good…or bad? Analytics centers of excellence are the sign of a mature organization…or they're just one of many potential options? Privacy concerns are something no one really cares about…or they are something everyone cares deeply about? Voices were raised. Light profanity was employed. Laughter ensued. This episode's Measurement Bite from show sponsor Recast is a brief explanation of statistical significance (and why shorthanding it is problematic…and why confidence intervals are often more practically useful in business than p-values) from Michael Kaminsky. For complete show notes, including links to items mentioned in this episode and a transcript of the show, visit the show page.

Data does not just magically spring into existence. Someone, somewhere, has to decide what data gets created and the rules for its creation. We would claim that this often starts as a pretty simple exercise, and then, over time, that simplicity balloons to be pretty complex! What if, for instance, you decided to listen to every #1 song on the Billboard Hot 100 going back to its inception in 1958? You may start by just capturing the song name, the artist, and the week(s) it was the #1 song. But, before you know it, you may find that you're adding in artist details…and songwriter details…and producer details…and genre details…and instrumentation details, and your dataset has 105 columns! But, oh, the questions that dataset could answer! And that's exactly the dataset that our guest for this episode, Chris Dalla Riva, created. He uses it (with a range of supplemental datasets) for his pieces in his Substack, Can't Get Much Higher, as well as the underlying raw material for his upcoming book, Uncharted Territory: What Numbers Tell Us about the Biggest Hit Songs and Ourselves. While the underlying material was music, the parallels to more staid business data were many when it comes to the underlying processes and challenges for doing that work! This episode's Measurement Bite from show sponsor Recast is an explanation of the miracle of randomization when it comes to addressing unobserved confounders from Michael Kaminsky! For complete show notes, including links to items mentioned in this episode and a transcript of the show, visit the show page.

If you didn't have a visceral reaction to the title for this episode, then you are almost certainly not in our target audience. There are few more certain ways to get a room full of analytics folk fired up than to raise the topic of dashboards. Are they where data goes to die, or are they the essential key to unlocking self-service access to actionable insights? Are they both? Is the question irrelevant, because, if they exist to inform business users, aren't they soon going to be replaced by an AI-powered chatbot, anyway? We thought a great way to dig into the topic (and, BTW, we were right) would be to have someone on the show who has co-penned multiple books on the topic. As luck would have it, Andy Cotgreave, one of the co-authors of both 2017's The Big Book of Dashboards: Visualizing Your Data Using Real-World Business Scenarios and the imminently releasing Dashboards That Deliver: How to Design, Develop, and Deploy Dashboards That Work agreed to join us for a lively chat on the topic! This episode's Measurement Bite from show sponsor Recast is a quick explanation of power analysis from Michael Kaminsky! For complete show notes, including links to items mentioned in this episode and a transcript of the show, visit the show page.

Imagine a world where business users simply fire up their analytics AI tool, ask for some insights, and get a clear and accurate response in return. That's the dream, isn't it? Is it just around the corner, or is it years away? Or is that vision embarrassingly misguided at its core? The very real humans who responded to our listener survey wanted to know where and how AI would be fitting into the analyst's toolkit, and, frankly, so do we! Maybe they (and you!) can fire up ol' Claude and ask it to analyze this episode with Juliana Jackson from the Standard Deviation podcast and Beyond the Mean Substack to find out! For complete show notes, including links to items mentioned in this episode and a transcript of the show, visit the show page.

Product managers for BI platforms have it easy. They "just" need to have the dev team build a tool that gives all types of users access to all of the data they should be allowed to see in a way that is quick, simple, and clear while preventing them from pulling data that can be misinterpreted. Of course, there are a lot of different types of users—from the C-level executive who wants ready access to high-level metrics all the way to the analyst or data scientist who wants to drop into a SQL flow state to everyone in between. And sometimes the tool needs to provide structured dashboards, while at other times it needs to be a mechanism for ad hoc analysis. Maybe the product manager's job is actually…impossible? Past Looker CAO and current Omni CEO Colin Zima joined this episode for a lively discussion on the subject! For complete show notes, including links to items mentioned in this episode and a transcript of the show, visit the show page.