Overview
For November, the Book of the Month is The Analytics Operating System: Teams, Governance, and the Path to AI by Fatima Soomro, Ph.D. A lot of books promise to fix your data program with a framework. This one takes a different route. It spins a fable, one that connects the challenges organizations have with data management in the world of AI.
Dr. Soomro says up front that she took inspiration from Patrick Lencioni’s The Five Dysfunctions of a Team. The book follows Bambylo, a fictional French toy retailer that built its brand on thoughtful stores and a toy buyback program. Then COVID hit and its neglected e-commerce site had to carry the business. The opening scene sets the tone. The CEO asks how many sales they made last month, and the answer is that it’ll take about a week to find out. Anyone who has worked in data has been in that room.
Part 1 tracks Bambylo’s data function from scratch. The team gets tired of gluing CSVs and data from a patchwork set of systems together. An IT team member named Amine gets promoted to Head of Data Warehouse. A new analytics lead, Gaurav, shows up and runs into reality. When the star schemas can’t keep up with more complex business questions, Gaurav brings in three pillars: Requirement Engineering, Data Engineering ,and Dashboard Engineering. The standout for me is his four-question intake. What’s the exact question? What decision will it enable? When is it needed, and why? Who will validate it? If a request can’t answer all four, the work doesn’t start. He also introduces “six data moves” (clean, aggregate, pivot, join, union, and scaffold) as a practical toolkit for analysts who’ve been jamming business logic into their dashboards.
Part 2 is where the book really hits home. The team grows, specialization turns into avoidance, and emergencies start setting the roadmap. One small team with no executive sponsor keeps getting pushed to the back of the queue. Their hunch that support tickets predict churn ends up killing Bambylo’s signature buyback program. Gaurav and Amine then inventory every dashboard and find duplicates, zombies, and the “dark matter” of critical spreadsheets living outside the official tools. Their answer is to sort analytics into Foundational, Operational & Tactical, and Exploratory & Strategic work. They wrap it in a layer of meta-analytics and set target time allocations. Soomro’s note sums up the section with a line everyone should remember:
“Wins don’t stay won.”
Part 3 takes on AI. A cost-focused CFO automates customer service, outsources payroll, and consolidates warehouses. Each project quietly pushes new work onto the data team. Eventually he asks the obvious question: why not automate analytics too? What follows is one of the most grounded takes on “chat with your data” I’ve read. The semantic layer pilot works well for Finance’s board pack, where the definitions were already agreed. It struggles in pricing, where nobody agrees on what success means. The book’s position is that dashboards work as contracts. They lock in definitions so groups of people can make recurring decisions the same way, and chat tools don’t replace that.
In conclusion, The Analytics Operating System is a delightful read for analytics leaders, CDOs, and the business executives who fund them. It’s especially useful for organizations in transition, the ones with some good foundations and growing AI pressure who still have time to get the structure right. It works as a story, and it also gives you frameworks and reflection questions you can take back to your own team. Ultimately, Dr. Soomro makes this an easy read that you can enjoy for the style just as much as the content, succeeding in its inspirational draw from Lencioni.
More About the Author
Fatima Soomro, Ph.D. is a data and analytics leader with over a decade of experience helping organisations turn analytical effort into effective decision-making. Her work draws on extensive hands-on practice and operational leadership, focusing on how analytics functions across teams, systems, and leadership to produce insights that translate into action.
