One Big Table (OBT) Pattern: Fan Traps, Grain Violations, and When Denormalization Actually Works
In Q3 of last year, a B2B SaaS company's growth team presented their quarterly board report. Marketing ROAS: 4.2x across all paid channels. LinkedIn ROAS: 6.1x. Google Ads: 3.8x. The board was delighted. Budget was reallocated toward LinkedIn. A senior marketer got a raise.
Eight months later, a new data engineer joined the team and did something nobody had done before: she reconciled the attribution dashboard against Stripe. Total attributed conversion value in the dashboard: $2.4 million. Actual revenue in Stripe for the same cohort: $880,000.
The discrepancy wasn't fraud. It wasn't a bug in Stripe or a misunderstanding of the attribution window. It was a fan trap, baked into the foundation of the marketing attribution One Big Table that had been powering every board report for three quarters.