SC 020 — Analytics Antenna & Good Guy Amazon?

Summary

Lays out a four-part analytics architecture plan for Squarely: Sources (Amazon KDP, Vercel, Fathom Analytics, ConvertKit), Ingestion (Fivetran, Steampipe, Meltano, custom Singer tap), Processing (DuckDB + MotherDuck with dbt-duckdb, S3 for raw storage), and Analysis (Evidence dashboards as weekly business reviews). The “antenna” metaphor comes from the founder’s father’s PhD research in signal error detection.

The closing section reframes Amazon as a force for good through KDP’s democratization of publishing — just-in-time printing, outsourced complexity, and near-zero barriers to entry. Contrasts with the founder’s failed TagaPet venture where wholesale inventory costs drained capital.

Key Arguments

• Start with dashboarding before A/B testing, operational analytics, or forecasting
• DuckDB + S3 is the right economics for a small business data stack — speed over granular access controls
• Amazon KDP eliminates inventory risk entirely: books print only after purchase
• Over 50% of Amazon retail revenue comes from third-party sellers — the infrastructure deserves credit alongside legitimate criticism

Writing Style Notes

Father’s PhD work as entry point is characteristic — personal stories frame technical concepts. The Amazon reframe is contrarian and grounded in lived experience (TagaPet failure). Practical tool choices reflect cost-consciousness appropriate to the business stage.

Connections

• 01-projects/squarely-puzzles/index — the analytics architecture is being built for Squarely
• 01-projects/newsletter/index — part of the Sanity Check body of work
• 06-reference/2026-04-03-data-maturity-processes-tools — choosing tools appropriate to maturity stage