Cashiu

**Problem:** Ember briefing reported "0 orders, 0 installs, 194 DAU" for Cashiu. The actual BQ data showed 50 orders, 34 installs, 200 DAU on Jun 8. **Root cause:** Two bugs: 1. `cashiu_daily.sql` used `COUNT(DISTINCT user_id)` on `ga_aggregated_event` but GA4 uses `user_pseudo_id` for DAU 2. `brief_json.py` looked for column names `myr_revenue`, `order_count`, `install_count` but SQL returned `cashiu_rev`, `orders`, `installs` **Fix:** 1. Updated SQL to use `event_name = 'purchase_item'` for orders and `user_pseudo_id` for DAU 2. Fixed column name mapping in `extract_cashiu()` **Impact:** Cashiu is NOT dead — it generates RM 1.3K-10K daily revenue with 35-130 orders and 200-281 DAU.

The May 19 briefing claimed Cashiu had "zero orders since May 13 (Day 5)" — this is incorrect as of live BQ data: - **purchase_item events** have been happening daily: 45 (May 13), 56 (May 14), 87 (May 15), 67 (May 16), 55 (May 17), 35 (May 18) - **validation_stats conversions** (with payouts): 39/10 validated/RM 59 (May 15), 142/17/RM 177 (May 16), 54/2/RM 84 (May 17), 75/0/RM 83 (May 18) - **DAU** for May 17 was 268 (session_start events), not ~170 as reported - **Signups** on May 17 = 0 (May 16 had 2 — the "2 signups" figure was from a prior day) The May 13-14 gap was real (no validated conversions those two dates) but the freeze resolved by May 15. Every day since May 15 has had purchase_item events, conversions, and cashback payouts. The "Day 5" (May 18) framing was stale — data shows 35 purchase_item events and 75 conversions on May 18 alone.

zero revenue/orders with engagement intact (50–59 installs, 163–176 DAU) — points to payment gateway or offer-configuration failure, not an app outage.

shipped 13 May 2026 with OpenReplay monitoring on Admin Dashboard and agent skill migration to Codex/Claude folders

collapsed ~56% on 11 May 2026 (MYR ~106 → MYR ~47) despite orders +15% and installs +50% DoD; cause undiagnosed, distinct from T+1 settlement-lag pattern

Each BU has its own performance metrics that the briefing should frame around, not generic terms like "throughput" or "collapse": - **Enterprise PG**: monthly initiatives per person (target 5/month, 60/year); new publisher sign-ups (target funnel: outreaches → pitches → onboarding → live); concentration risk - **Enterprise AG**: advertiser revenue split per account manager; offer health decay; named advertiser deals - **Subscriptions**: active account count, MRR concentration by client, integration request queue, pitch tracker - **iAds**: ROI%, offer count, publisher count, named partners (Trip, Stars, JP) - **CMV Video**: ClickUp deal pipeline (Leads → Proposal → Confirmed), billing AR, brand milestones, smartphone launch coverage - **Cashiu**: installs, DAU, signups, AOV, daily orders, daily revenue - **CMV Affiliate**: P&L by month, key publishers - **Group Services**: headcount, staff costs, marketing budget burn, payment/billing queue - **Corp Dev**: MDEC approval status, exit process milestones (iPrice deal is CLOSED — see CORRECTION 2026-06-01) Briefings should cite the SPECIFIC named publishers/advertisers/clients/initiatives — not abstract metrics. "Mind has 59 pitches, only UC Browser converted" beats "PG throughput collapsed".

Brief output must use plain operator language. Specifically: - **CVE** → "security hole" or "security vulnerability". Always name what the affected library/component is and what the practical risk is ("code execution if attacker hits URL X", "data leak if file Y is requested"). dompdf/swiper/chart.js mean nothing to Jim — say "the PDF generator", "the UI carousel library", "the chart library". - **CPS/CPI/CPM/CPC** → "commission per sale / install / impression / click" on first reference per item, then OK to abbreviate. - **T+1 / T+2** → "yesterday partial / two-day-ago final" - **DoD / WoW / MoM** → "day-over-day / week-over-week / month-over-month" on first use - **AOV** → "average order value (revenue ÷ orders)" - **ROI / ROAS / GP** → first-use unpack - Internal product codenames (4223, 5126, INVVM-1721) — keep the code but always add a parenthetical noun ("invoice INVVM-1721", "Cashiu attribution card 4223") Rule of thumb: if Jim can't act on the item without first decoding a term, the term is a bug. Translate or unpack.

When the brief flags a numerical anomaly (X "halved", Y "collapsed", Z "outlier"), it must check the trailing distribution before claiming significance: - Pull the last 7-14 day series from the relevant BQ metric - Compute mean / range / std-deviation - State whether the flagged value is genuinely anomalous (>2σ, outside band) or within normal volatility - If anomalous, propose 2-3 specific hypotheses (settlement lag, mix shift, prior-day outlier, attribution residue) — NOT generic "undiagnosed" Example of what NOT to do (today's brief): - "Cashiu AOV halved May 10→11, undiagnosed" The actual data showed: trailing 7d AOV range 33-106, mean 53, std-dev ~25. May 11 at 46 is WITHIN normal range. May 10 at 106 was the outlier needing explanation, not May 11. The brief anchored on the largest delta instead of running the distribution. With BQ data in hand the brief can and should do this analysis itself.

There is a publisher account named "Cashiu" on the Involve Asia affiliate platform (publisher='Cashiu', affiliate_id=1016390). This generates ~MYR 170/day on average over 15 days observed — also small. DO NOT conflate the two: - **Cashiu the product** — Commerce Media Ventures' cashback app. Its KPIs (installs, orders, DAU, revenue) come from `productnation-data-analytic.cashiu.*` tables. Real scale: dozens of installs/day. - **"Cashiu" the publisher** — an affiliate account on Involve Asia's network. Generates ~MYR 170/day in commissions. Appears in `jimmy.conversion_agg_offfer_promotion_country` with affiliate_id 1016390. Different thing. When the brief says "Cashiu revenue X" it must be unambiguous about which is meant.

A 1,000,000× inflation bug went undetected for at least a week because no brief reviewer (human or model) sanity-checked the order of magnitude. Rule: before reporting any single-day Cashiu/product/BU number, ask whether the magnitude is plausible for that product's scale. Cashiu installs in the millions per day would imply a Tier-1 app — clearly not the case. The brief should flag implausible numbers ("Cashiu reporting 32M installs/day — likely a data join issue, not actual scale") instead of citing them.

32.7M installs against zero revenue/orders, unreconciled — qualitatively different from prior T+1 settlement lag pattern

shipped 7 May 2026 (Khairul); ten-day overdue flag resolved

accepted as operating model across Cashiu (Hanh), iAds (Tay/Tan), CMV Video (Andrew Tai), Subscriptions (Susanne ~60%) — no longer surfaced as flag

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

confirmed structural across May 4 and May 5 (MYR 105.9M and MYR 189.5M back-fills); P&L regressed from 848 MYR to 77 MYR same cycle, unexplained

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

zero revenue, zero orders posted — cause unresolved as of 05 May 2026

shipped 29 April 2026, introducing the Publisher Account Base Figures setting

cashiu-mobile-1.25.0 two days slipped past 30 April with no revised ETA