Andrew Kritzer

Most growth-stage companies don’t have a data problem. They have a leadership and clarity problem. My conversation with Ryan Koonce on Health Tech Builders put sharper language around a pattern I see often: Being “data-driven” is not about dashboards, metrics, tools, or hiring more analysts. It’s about leaders deciding—explicitly—what data is for. Ryan sees the same issues across company after company: Exec teams use the same metrics with different definitions Google Analytics is treated as decision-grade (it’s not) Leaders ask dozens of questions, but prioritize none Data teams become bottlenecks instead of accelerators As Ryan put it: “Data isn’t just the number. It’s the definition behind the number.” And this one really resonated: “If you’re not looking at the most important data, it doesn’t matter how accurate it is.” The companies that actually turn data into growth do a few things differently: CEOs own metric definitions Canonical metrics are documented and standardized Clean data foundations are built early (data warehouse + CDP) Google Analytics is treated as directional, not decision-grade Leaders focus on a small number of business-critical questions Performance is benchmarked internally, not externally Executives develop data literacy, increasingly using AI directly If you’re a CEO or operator wondering why analytics hasn’t translated into better decisions, this episode is worth your time. Link to the full interview in the comments. #Leadership #DataStrategy #HealthTech

7

2 Comments

AstroBee generates ontologies...but what is an ontology, and if they're important, why don't you already have one? I've written a short piece on Kimball star schemas, ontologies, and how each relate to Bill Inmon's original approach to data warehousing in the early 90s. Dig in to find an overview of a fascinating technological debate and the re-emergence of an approach to data modeling that's been sleeping for 20 years.

18

3 Comments

Most metrics never lead to action. Over the past year, I’ve been building something that changes that. I call it Compass. It started with a simple observation: We’re very good at building dashboards. But cross-functional teams don’t always have a clear path from insight to decision. This isn’t a tooling problem. It’s an operating gap. I’ve seen it play out repeatedly: 🔴 The KPI turns red ❓ The team debates whether the metric is right 💬 Conversations expand 📁 Request for approvals stack up ⏳ Action slips from days to weeks The issue isn’t visibility. It’s the missing bridge between “what we’re seeing” to “what we’re doing next.” When that next step isn’t defined: ⏳ time drifts 🦥 momentum slows 📉 performance discussions expand instead of converge But when the next action is clear — ownership, response, pilot — the metric stops being a report and starts becoming a driver of strategy. That’s what Compass is designed to solve. Compass is a decision-driven KPI framework that connects what’s usually scattered across tools and meetings: • Outcomes — the goals teams are trying to achieve • Drivers — what influences those outcomes • Metrics — how progress is measured • Decisions — what leaders choose to do next • Pilots — the initiatives that test and improve performance It gives teams a clear system for turning insight into action — not just reporting on it. And it’s not just for analytics teams. Compass is built for the people responsible for making strategy happen: • Operations • Program management • Analytics • Sales • Marketing • Finance From startups to enterprise organizations — if you’re accountable for outcomes, this is for you. Two things people said immediately when they saw it: 1️⃣ “This isn’t just for analytics — this is for everyone.” Operators and team leads can act on metrics without needing to be analytically fluent. 2️⃣ “It connects the metric to what I need to do next.” Ownership, pre-aligned responses, and structured pilots — built in. By reducing the coordination overhead around metrics, Compass helps leaders move from reporting performance to actively shaping it. This week I’ll share a short series breaking down how Compass works: 1️⃣ Analytics → Decisions 2️⃣ Pilots → Execution discipline 3️⃣ The Daily Brief → Operational focus If you’re working with leadership teams trying to turn strategy into consistent execution, I’d value an introduction. Compass isn’t another reporting layer — it’s an operating framework for turning numbers into motion.

17

4 Comments

Every data team knows the whiplash of a foundational initiative. You spend a year lobbying for a beautiful foundational layer. You finally get it greenlit for the next fiscal. You build it for six months. Mid-year, the CEO asks if you can pivot to “this other urgent business thing.” You can’t explain why you can’t. The pivot gets prioritized. Development stops. Entropy takes your half-built foundational layer to a better place beyond this world 🪦 Data teams: please don’t let this happen to Retrieval. We're all about to get that 2026 approval for foundational knowledge management. Please do things differently and protect your beautiful knowledge graph, ontology, semantic layer investments. How: 🌟 North Star → Get crystal clear on what you’re ultimately building. If you haven't written a white paper, you’re not clear. 🏎️ Fast Start → Pick a small, high-ROI use case that forces you one step toward the North Star. Ship only that - not the whole platform! Go fast. Measure value. 📣 Loud Part → Communicate success like your life depends on it. To your team. To your boss. To your CEO. Over and over. Beat your CEO to the punch - they should know about the ROI before they get those mid-year ideas. Repeat for the next use case. Shipping should get easier as foundation becomes sturdier. At Salesforce, our first use case was lead routing. It made millions. We’re not done - but now our focus is on shipping, not fighting entropy. h/t Miklós Molnár for talking about the power of fast high ROI use case

48

10 Comments

T-1 week to #ReGovern2025!  We've been talking to hundreds of data leaders this year, and one thing's clear: there is no playbook for the AI era.The best data teams aren't waiting for one either. They're writing it- 12 months ahead of everyone else. And next week, they're showing up to share what they've learned. Join the conversation: https://lnkd.in/ep6eTTWw

17

🧵 Stitch Fix: Elevating the Art + Science of Styling with Knowledge Graphs For years, Stitch Fix has been the industry benchmark for blending human intuition with data science. But in a world of rapidly evolving AI, how do you stay ahead of the curve and continue to inspire individual style journeys? I had the pleasure of sitting down with Meg Paulosky, Director of Data and Analytics at Stitch Fix, at the last Open Data Science Conference (ODSC) to discuss how they are using graphs to move beyond traditional data models. "Exploring not only AI, but knowledge graphs has really helped us reimagine what [the style journey] looks like in a way that works for everybody... " Three Takeaways from our Conversation: 1️⃣ Architectural Innovation: Shifting the focus from basic data points to sophisticated Agent Design and systems modeling. 2️⃣ Universal Personalization: How graphs allow Stitch Fix to scale style inspiration beyond their existing clients and services. 3️⃣ The ODSC Pulse: Why it's essential to meet other innovators and see tangible, cross-industry AI applications in person. It was incredible to get a few minutes with such a forward-thinking leader at one of the best data science events. (Next ODSC is in April!) 🎧 Listen to the Graph Chat here: https://lnkd.in/etUqtVWW 🎬 Watch here: https://lnkd.in/ePxSzutJ

42

2 Comments

Congrats to the team at Nimble! We've been using their scraping capabilities for years (they have some of the best Social scraping capabilities out there. Excited to see this move towards agentic search. Something I didn't get for some time was basically - "Why do agents need websearch... doesn't pretty much every AI now have websearch abilities anyways?" Think: I could already chat with my openclaw or an agent I built on claude code and tell it 'enrich all these companies with their respective www or no. of employees' BUT it takes a long time. And that's the tell. It's slow because your agent is doing what a human would do... googling each company, reading the page, summarizing what it found, moving to the next one. It's getting answers. Not data. The difference: Built-in AI search returns prose your agent has to interpret. Nimble's API returns structured JSON your agent can immediately act on — write to a database, pass to the next step, trigger a workflow. No interpreting. No parsing. No waiting for the model to "read" a webpage. And because Nimble is running real headless browsers against live sites (not querying a cached index), the data is accurate as of right now — not last week. Ilan Chemla just published a great walkthrough of exactly this: a company enrichment agent using Nimble + Databricks + LangChain + Claude Sonnet https://lnkd.in/g4mNJd6V that processes thousands of companies at scale. What your agent does slowly in a chat window, this pipeline does in parallel across a cluster. Powerful an capable of enterprise grade use cases and datapipelines but also for vibe coders like me: you're basically able to just give the api to your agent (do it safely) and tell it to remember that now it has that capability to search the web WAY better.

5

Airtable's CEO asked one question and restructured his entire company around the answer: "If we founded this business today with AI as a core capability, how would we operate?" The answer was two speeds with fast-thinking teams that experiment weekly and tolerate mess and slow-thinking teams that build durable infrastructure with precision. Most companies try to do both with one team, one process, and one cadence. The experimental work dies under process weight, or the core business gets unstable from moving too fast. The insight here isn't that these aren't execution failures but structural ones. You can't move fast and slow at the same time with the same team. A two-speed model recognizes that experimentation and infrastructure are fundamentally different work. Forcing them through the same process is a design flaw. #AIStrategy #OrganizationalDesign #AITransformation #Leadership

4

7 Comments

Returning 30 million clinical minutes to healthcare providers using open-source models in production is incredible. That is not a leaderboard stat. That is real time back with patients. Less admin. Less burnout. More care. A couple of things stand out to me: Open models drive real results. When paired with the right inference infrastructure and deep domain expertise, they become powerful, customizable tools that solve real problems. And measuring the outcome you actually care about up front changes everything. When the goal is clinical minutes returned, not just model accuracy, you build differently. You deploy differently. You prioritize differently. Huge congrats to Sully.ai and Baseten for showing what this looks like in the real world.

31

1 Comment