Data Driven Decision Making

Part 0 · Chapter 0

From Data Traces to Decisions

Treat data as a trace of business activity, then route the question through storage, evidence, and a decision someone can own.

This chapter draws the map the rest of the book follows. Its argument is that a manager's missing skill is not another technique but an operating discipline — one that links a business question to the data trace that records it, the storage system that holds it, and the decision someone has to own. Following a single Bean & Basket customer's Tuesday morning, it shows how a search, an impression, a transaction, a four-star review, and an AI log capture different slices of behavior with different blind spots, then sorts the storage stack (operational SQL, warehouses, local analytics, vector and graph stores) and the workflow families that turn evidence into action. It closes on the data-to-decision loop, where genuine data-driven work is separated from data-decorated work by three ingredients: an action, a counterfactual comparison, and a threshold.

Start reading

Topics covered

activity bias and workflow biasdata trace vs. truthtransactional vs. analytical systemswarehouses, lakes, and local analyticsvector and graph storesbatch vs. streaming freshnessthe use-case routerthe data-to-decision loopdata-driven vs. data-decorated decisionsmetric, model, and AI workflow cards

In this chapter

  1. 0.0Foreword: How to Read This BookLays out the book's wager, audience, reading paths, the Bean & Basket through-line, and the standalone cases that ground later parts.
  2. 0.1Where Data Comes FromReframes every dataset as a trace of specific business activity, with a customer-morning example and three generation traps to guard against.
  3. 0.2How Data Is StoredMaps the storage stack by job, separating transactional from analytical systems and matching data freshness to the cadence of the decision.
  4. 0.3How Data Is UsedSorts data work into recurring workflow families and routes a business question to monitoring, diagnosis, causal, prediction, or AI evidence.
  5. 0.4The Data-to-Decision LoopConnects source, storage, evidence, action, and feedback into one loop, distinguishing data-driven decisions from merely data-decorated ones.