# Desk log EAT-JOIN-20260823

**Status:** First-party illustrative retail composite; not customer, order, vendor-comparison, accounting, benchmark, or independently audited data  
**Run date:** 2026-08-23  
**Evidence checked:** 2026-08-31  
**Operator:** InfiniSynapse Data Team  
**Attestor:** William Zhu, InfiniSynapse cofounder ([GitHub @allwefantasy](https://github.com/allwefantasy))

The operator compared an illustrative session rank with an order/SKU/cost join over 18,400 lines. The output displayed 12 session-ranked SKUs, four without dated cost, six return-lag flips, two additional fee flips, and a 9.4% return rate for the flip set. Paid spend and a reorder list were held because the output lacked causal evidence and complete dated cost.

The [CSV](https://infinisynapse.com/blog-media/ecommerce-analytics-tools/downloads/tool-output-EAT-JOIN-20260823.csv) exposes six aggregates and two held actions. The [verification script](https://infinisynapse.com/blog-media/ecommerce-analytics-tools/downloads/verify-EAT-JOIN-20260823.py) checks displayed output only. Source rows, SKU identities, monetary values, SQL, cost vintages, vendor configurations, denominators, and reconciliations are unavailable.

External context: [U.S. GAO Assessing Data Reliability](https://www.gao.gov/products/gao-20-283g), [UK Government AQuA Book](https://www.gov.uk/guidance/the-aqua-book), and [U.S. Census Quarterly Retail E-Commerce methodology](https://www.census.gov/retail/ecommerce/how_surveys_are_collected.html) (retrieved 2026-08-31). These publishers did not review this run.