dataeffet
← Back to All Technical Case Studies
Enterprise Portfolio$24.6M Annual Run Rate (NA & EU)310+ Active SKUs

Architecting a Sovereign BigQuery & dbt Lakehouse to Replace SaaS Middleware Across a $24.6M Amazon Portfolio

OmniPulse was paying $4,200/month across third-party analytics SaaS platforms while struggling with 48-hour data latency. By migrating to a sovereign Dataeffet OS Medallion pipeline in their own Google Cloud environment, they eliminated SaaS subscriptions entirely, reconciled daily P&L down to the carton level, and reduced TACoS by 4.2% within 90 days.

SaaS Middleware Savings
$50,400 / yr
Data Processing Latency
48 Hrs → Real-Time
True TACoS Improvement
-4.2% Absolute
Excel Macros Required
0 (Zero)
The SaaS Middleware Bottleneck

Why SaaS Aggregators Failed OmniPulse Consumer Hardware

Operating 310+ SKUs across North America and Europe created an intractable data bottleneck. Third-party SaaS tools aggregated ad spend and inventory positions via black-box ETLs that failed to reconcile with Amazon settlement reports, leaving finance teams to manually audit variances in 800,000-row spreadsheets.

Consequence: Leadership was paying $50,400/year for third-party analytics dashboards while making inventory reorder decisions based on 48-hour-old data.
Sovereign Lakehouse Architecture

How Dataeffet OS Reclaimed Margin

Deployed a client-owned 4-Tier Medallion pipeline (Bronze, Silver, Gold, Diamond) utilizing dbt Core and BigQuery. Raw SP-API settlement ledgers, AMC ad paths, and Freightos container rates are now ingested directly and calculated deterministically without middleware latency.

Result: Client-owned BigQuery warehouse providing automated, deterministic daily net margins and True TACoS without macros.

Pipeline Architecture & SQL Teardown

OmniPulse OS • BigQuery dbt Medallion Audit Engine
// Step 1: dbt Silver Transformation Model (Automated Daily Schedule)

SELECT sku, SUM(amazon_settlement_payout) - SUM(landed_cogs_usd) - SUM(attributed_ppc_spend) AS absolute_net_margin_usd FROM gold_omnipulse.daily_sku_economics WHERE settlement_status = "RECONCILED" GROUP BY 1 ORDER BY 2 DESC;

-- Output guarantees zero unmapped SP-API transactions
-- Automatically allocates landed ocean freight & 3PL prep costs per ASIN
01. Bronze LayerRaw JSON ledgers ingested from Selling Partner API & Ads API.
02. Silver & Gold Layersdbt models clean data and hardcode landed freight and PPC waste rules.
03. Diamond LayerExecutive AI query engine accessible via natural language terminal.

Build Your Brand's Sovereign
Medallion Data Infrastructure.

Deploy the OS directly into your client-owned BigQuery environment. Regain absolute control over your Amazon SKU unit economics today.