Three Google products share the Looker name. Two are products. One is a language. All three do fundamentally different things, serve different audiences, and cost very different amounts. This article maps them clearly — including when they work together.
Read Article →The most complete guide to Looker Studio (formerly Google Data Studio): interface walkthrough, all connector types, every chart, calculated fields with real formulas, data blending step by step, parameters, pro tips, honest limitations, and a detailed comparison with Power BI and Tableau.
Read Article →30% of your SFCC orders have no GA4 session. This is not a bug. It is a mix of five structurally different categories — each requiring a different response. This guide covers every category in depth, the compounding analytics errors they cause if ignored, a two-tier LookML classification system, all the metrics to build on top, and the complete reconciliation dashboard architecture.
Read Article →Five dashboards, five revenue definitions, one team. This is not a data quality problem — it is a governance problem. This guide covers the full LookML governance approach: the right way to define dimensions and measures, cross-source metrics, derived tables, Git versioning, access control, and the real-world workflow that prevents metric drift permanently.
Read Article →Your LEFT JOIN between GA4 and SFCC returns 40% NULLs. The problem is almost always the transaction ID. This article covers every failure mode with real examples, a decision-tree for diagnosing your specific case, four SQL queries to run before touching Looker, three LookML fix patterns, and a benchmark table for what healthy match rates look like by segment.
Read Article →GA4, SFCC and your offline files will always show different revenue numbers. This is not a bug. Each system measures a different reality, at a different moment, with a different definition of what counts. This deep-dive covers every structural cause, the full data flow architecture, practical BigQuery diagnostics, and the business questions that only become answerable once you understand why the numbers differ.
Read Article →Looker is not a dashboard tool. It is a semantic layer platform built around LookML — a modeling language that defines your business logic once and enforces it across every report, every user, and every team. This guide covers everything: architecture, LookML fundamentals, joins, calculated metrics, derived tables, access control, scheduling, performance, and real-world use cases.
Read Article →Introduction Server-side tracking is often presented as a universal fix for: The common belief is simple: “If tracking runs on the server, browsers can’t interfere.” This is incomplete. Browsers do not block based on where your container runs.They apply restrictions based on: Server-side tracking is not a shield.It is an architectural shift. And architecture can […]
Read Article →Cookieless is not determinist Clearing cookies used to feel like a reset. It is not. Modern tracking increasingly relies on measurement (fingerprinting) and side-channels (caches and protocols people rarely think about). To make this concrete, I tested my own browser with the EFF’s Cover Your Tracks tool. The result was a unique fingerprint among 303,650 […]
Read Article →Introduction The term “cookieless” has become the hottest buzzword in digital marketing and ad tech. You can’t attend a conference, read a vendor pitch, or scroll through LinkedIn without seeing someone touting their “cookieless solution.” Google keeps postponing the deprecation of third-party cookies in Chrome. Safari and Firefox have already blocked them. The GDPR and […]
Read Article →Three containers. Perfect architecture on paper. 60% unset variables in production. That was my week. A client approached me with what seemed like a textbook Tag Management System setup: one core container for essential tracking, two business-unit-specific containers for specialized marketing tags. Clean separation of concerns. Different teams managing their own implementations. Excellent governance model. […]
Read Article →You’ve spent hours configuring Google Tag Manager. Every variable is mapped. Every trigger fires perfectly in Preview mode. You test, retest, and finally hit publish with confidence. The next morning, you open GA4 reports. 40% of your custom event parameters show “(unset)”. Welcome to one of the most frustrating debugging experiences in web analytics: the […]
Read Article →