BigQuery

From Lan Zhi Xiong's Friday Tech WIP (May 15), surfaced directly by Jim in Telegram. The briefing never surfaced this properly because Management Updates project (39206207) routes to Leadership, not Engineering: **TikTok Data Integration Crisis (CRITICAL):** - TikTok = 20% of total revenue via 18 publishers - Major data accuracy issues: duplicate transactions, zero amounts, partial trust infections - Required company-wide cleanup effort lasting ~1 month - Root causes: no staging environment, insufficient monitoring, API unreliability forced switch to dashboard scraping - Fixes planned: staging env (1-2 week testing), code-level guardrails, engineering ownership of dashboard automation **Shopper platform:** - Still generating ~4-figure revenue with active users despite planned exit - Decision: move to subdomain with single landing page for cashback, not kill immediately - Target: end of Thu/Fri, internal testing next week **Weekly deployment cadence already broken:** - Missing this week due to B3 tag 4 migration complexity - UI components break when clicked despite looking correct - Advertiser dashboard: 1 onboarded user, low confidence in active usage **Security gaps ahead of audit:** - Credential management issues (passwords not in bot manager) - Need service account strategy for BigQuery and other tools - API access controls need review (self-approval currently possible) **Next Steps flagged:** - Jimmy: review security access, establish service account policies - Jimmy: get product team update from Osama (30 min weekly check-ins needed) The Tech WIP posts are in Management Updates project (39206207) under `[Tech]` tags but route to Leadership synthesis — they need to feed Engineering synthesis instead.

MYR 98.4M / 1.12M orders — third consecutive T+1 back-fill, settlement lag now confirmed structural

MYR 105.9M / 1.4M orders, attributed to settlement validation lag; resolves the zero-revenue flag from 05 May 2026

MYR 105.9M daily BigQuery vs MYR 848 May net P&L — order-of-magnitude irreconcilable as of 06 May 2026

Daily BigQuery revenue figures fill in over ~48 hours post-close as conversions validate. The "Apr 30 three-way split" (MYR 262K / 494K / 542K read on consecutive days) was Ember reading the same date as lag resolved — not three conflicting numbers, just one number landing in three reads. **Stop flagging day-by-day revenue figure changes as a "pipeline integrity crisis."** Daily revenue actually was stable on Apr 30 per Jim's direct read of the system.

10.9% of network revenue as of 30 April BigQuery extract