Analytics, BI & decision intelligence
Deliver dashboards, self-serve and advanced analytics to build a data-driven culture
Analytics is where most organizations lie to themselves. They buy dashboards nobody reads, celebrate vanity metrics, and confuse having data with making decisions. This block fixes that. It takes you from modern BI foundations all the way to models running in production, and it treats analytics as what it actually is, a decision engine for the business.
You start with the fundamentals that separate mature data organizations from the rest. Modern BI tooling, the semantic layer that keeps everyone speaking the same language, dashboard design built around KPIs that actually drive action, and self-serve analytics backed by a real data catalog and genuine data literacy across your teams.
Then you move up the value chain. Decision intelligence reframes analytics around the architecture of decisions themselves, with embedded analytics putting insight exactly where the work happens. You get into advanced techniques that finance and the board respect, customer lifetime value, churn prediction, and demand forecasting. And because none of this matters without proof, you learn to build the business case and defend the ROI of every analytics investment you make.
Finally, you tackle the part everyone underestimates. Experimentation at scale, done with real causal rigor rather than A/B theater, and the discipline of running models in production. Drift, monitoring, and MLOps are not engineering trivia, they are the difference between a model that creates value and one that quietly rots.
By the end you will stop treating analytics as a reporting function and start running it as a competitive weapon. That is the standard a CDO is held to, and this block gets you there.
What you'll master
- Design dashboards and KPIs that drive decisions instead of decorating meetings
- Build a self-serve analytics capability with a data catalog and real data literacy
- Architect decisions using decision intelligence and embedded analytics
- Deploy advanced models for CLV, churn prediction, and demand forecasting
- Prove the business value of analytics with a rigorous ROI and business case
- Run experimentation at scale with genuine causal rigor
- Keep production models healthy with drift detection, monitoring, and MLOps
Modules
Build modern BI and self-serve analytics with a semantic layer and data catalog
Build advanced analytics and decision intelligence use cases while measuring their ROI
Industrialize experimentation and model deployment to production with MLOps
Shows how to turn dashboards into decisions, craft data narratives for executives, and embed analytics directly into daily workflows.
Shows how to define metrics in a semantic layer, build KPI trees, and set north-star and guardrail measures that align decisions.
Frequently asked questions
What does the Analytics, BI & decision intelligence block actually cover?
It covers the full chain from modern BI foundations to models running in production: semantic layers, dashboard and KPI design, self-serve analytics, decision intelligence, CLV, churn prediction, demand forecasting, experimentation and MLOps. The block is organized into 5 modules of 3 lessons each, 15 lessons in total. The framing throughout is analytics as a decision engine, not a reporting function.
Who is this for, and do I need to code?
It is written for the person accountable for the data function, typically a CDO or a data leader, and it sits inside the CDO Track. No coding is required: the material deals with architecture choices, metric definitions, ROI arguments and operating discipline rather than syntax. You should be comfortable reading a KPI definition and challenging a model's business case.
What is the difference between business intelligence and decision intelligence?
Business intelligence produces reports and dashboards that describe what happened; decision intelligence starts from the decision itself and organizes data, models and workflows around it. In practice that means mapping who decides what, on which cadence, then using embedded analytics to put the insight inside the tool where the work happens. The block treats BI as the foundation and decision intelligence as the layer that makes it useful.
Where should I start if nobody reads our dashboards?
Start with the metrics and semantic layer before touching the dashboards. Unread dashboards are usually a symptom of metrics that mean different things to different teams, or KPIs that no one is accountable for acting on. The lessons on the semantic layer, KPI trees, and north-star and guardrail metrics address that root cause; dashboard design only pays off once the definitions are settled.
Do I need a data catalog before opening up self-serve analytics?
Yes, or self-serve becomes a factory of contradictory numbers. The lesson on self-serve analytics pairs the architecture with a data catalog and real data literacy across teams, because giving people query access without documented, trusted definitions shifts the reporting bottleneck into a credibility problem. The catalog is what lets a business user find the right table and know who owns it.
What does the block say about keeping models alive once they are in production?
Two of the fifteen lessons are devoted to it: one on models in production covering drift and monitoring, and one on MLOps covering monitoring, retraining and drift. The argument is that a model without drift detection degrades silently, so the value it created erodes without anyone noticing. Monitoring and retraining cadence are treated as management decisions, not engineering details.