Ministry of Oats · Internal

Org Design —
Assessment & Proposal

Proposal · not yet decided

A read of where the team's actual JDs and workload sit today, against a 60‑day window of real operating channels — not a re-statement of intent, an assessment of what's actually happening on the floor. Ends in a proposed program map and per-person KPIs, both open for debate.

The one-line finding

The platform build and the org build were supposed to co-evolve. Neither half delivered — and that's not a coincidence. There was never an up-leveled operator role ready to use a dashboard-style tool, so tooling built for that role had nothing to land on. The org gap and the tech gap are the same gap, not two separate problems.

Where this comes from

Not inference from job descriptions or memory — a direct read of the channels where the business actually runs, cross-referenced against three independent sources so no single one is trusted alone.

April JDs vs. August ground truth

Three JDs were drafted mid-April with Q1 asks attached. Checked each against what the 60-day window actually shows.

PersonJD's Q1 askWhat the window showsVerdict
Abhi Own category + aggregator P&L end-to-end; Pratik exits the Tuesday table by Jun 30 Operating independently on Cart + Category — Zomato tier negotiation, CPG box print approval, procurement escalations, all self-directed Mostly held
Manju Stop being kitchen ops' single point of failure; 2 promoted store leaders own daily ops by Jun 30; Manju supervises from central, walks stores weekly not daily Did not happen. One person is still the daily RCA-enforcer, KPT-chaser and discipline-issuer across 4 separate WhatsApp channels — exactly the single-point-of-failure pattern the JD tried to kill Failed — the single biggest org gap
Rushil MOO at X partner sales live, 3K embedded orders/mo exit rate, by Jun 30 Breakfast & Business was "yet to start" per his own July marketing deck; first real Fridge placement (Goodera) lands Aug 31 ~2 quarters late

Manju's own JD line: "10% — Outpost Platform adoption. Drive store module to 100% compliance once it ships." Neither half delivered, on the same person's JD — reinforcing that this isn't two separate problems.

Notes from the floor

Specific, dated evidence behind the assessment above — not summary claims.

RCA quality — never fixed, named 3× by the ops lead himself

"RCAs are done to stop the complaints not to repeat the same thing again and again"MOO Leaders · Jul 4
"Guys why the RCAs not shared yet ?"MOO Leaders · Aug 13

Koramangala ratings crisis

"I'm really concerned on the ratings of Koramangala… For the next 15 days just focus on this… I'm okay if the KPT is high but not with complaints and low ratings."CORE OPS Outpost · Aug 21 — 15-day remediation plan, outcome unconfirmed as of window's end

The KPT vs. quality tradeoff, named explicitly

"I'm okay with higher KPT but not the missing and wrong order"MOO Leaders · Aug 5

Blueberries — flagged urgent 4 times in one month

Central Store stock-out escalations on the same item: Aug 18, Aug 20, Aug 22, Aug 26 — the one recurring shortage that never got durably fixed across the whole 60-day window.

Week 29 COGS by store (Jul 21, shared in-channel)

StoreFood costPackagingOverall COGS
Whitefield20.86%4.89%26.31%
Indiranagar23.31%5.43%29.40% (flagged as inflated by a Week-28 inventory error)
HSR20.91%5.43%26.34%
Overall21.77%5.29%27.44%

From Granola — the arc of the last 6 months

Mar–Apr: pure store firefighting → Jun–Jul: category/product strategy → Aug: almost entirely fundraise. Matches the WhatsApp read exactly: ops has been running on autopilot while attention moved upstream. Two recurring items that never closed: the JP Nagar 40%+ COGS mystery ran for months before being traced to inventory entry error, not real shrinkage; a ₹5,000/store staffing incentive existed for months with zero stores ever qualifying for it.

Proposed program × owner map

Not "7 programs, pick an owner" — spread the load off Abhi and onto Manju's now-real bench, and use Ankita as the incubator for anything genuinely unowned.

ProgramOwnerWhy
CartAbhi unchangedStrongest, most JD-aligned work — don't touch it.
Kitchen OpsManju (oversight) → Yiben + Rishabh (execution)The delegation April promised and never happened — now there's an actual bench to hand it to.
DispatchManju (oversight) → Shivaraj (execution)Central kitchen production + dispatch is literally his job description.
Breakfast & BusinessRushilAlready has real motion (Neighbour.com DIY-bar deal, Goodera Fridge) — needs clean KPIs and a reporting rhythm, not a new muscle.
Oatmeal Gang (membership/community)Ankita, new-initiative stand-upGenuinely unowned today. Not handed to Rushil on top of B&B — that repeats the Abhi mistake.
CPGAnkita, new-initiative stand-upReal motion (packaging already printing) with nobody accountable end-to-end. Abhi is explicitly overloaded — don't add to him.
BrandMaintained function, no owner-JD yetRunning fine at low intensity via Contentlogy + occasional Rushil oversight. Don't manufacture a program for something that isn't broken.

Abhi — the logical binding

His current 5 JTBDs, split into keep vs. move. The overload isn't a capacity problem — 3 of 5 responsibilities were never really "Cart," they got absorbed by default because nobody else existed to catch them.

JTBDVerdictWhere it goes
Discounting/ads across platforms + category P&LKeepThis is Cart, his core — automate the daily ritual harder
Category strength (existing menu performance)KeepOther half of the same weekly review he already runs well
ProcurementMoveA supply-chain function, not category — goes to Manju/Shivaraj's Dispatch
New recipes / NPDMove (mostly)Recipe Room already exists for this — route through CPG owner, Abhi consulted not owning
New category launchesMoveSame logic as NPD — a CPG job, not a weekly-Cart-review job

Automation opportunities — Manju's bench, via the Rista API

  1. Variance — same-day entry-vs-system mismatch detection instead of a mid-month mystery (the JP Nagar case, solved).
  2. GRN/dispatch — currently 100% WhatsApp screenshots. Rista already has the transaction data; this is a reporting gap, not a data gap.
  3. COGS-by-store, weekly — already computed by hand every week from data Rista already has. Close to a one-click report.
  4. Supplier scorecard — a standing view would have caught MFP's ~₹42K balance + repeat quality complaint before it sat unresolved.

Shivaraj is the natural first user of #1/#2/#4; a store-level version of #1/#3 is what Yiben/Rishabh's new store-audit scope needs.

Per-person KPI + cadence — starting point

Proposals to pressure-test with each person directly, not final until they've had input.

PersonCore KPIs (proposed)Cadence
AbhiBlended aggregator margin, AOV, WoW store growth (3 stores), listing-health incidents/weekDaily numbers ping (automated) · weekly Thu review · monthly P&L close
RushilB&B pipeline by stage, events closed/mo, revenue/event, Fridge placements liveDaily pipeline pulse · weekly sprint review · monthly report
ManjuDirect margin %, critical QC incidents, % of daily ops handled by Yiben/Rishabh (the real handoff metric)Weekly store-leader 1:1s · weekly rotating store walk · monthly leadership review
ShivarajDispatch on-time %, production variance (planned vs actual), GRN turnaroundDaily central priority · weekly logistics sync with Manju
Yiben / RishabhStore audit score, KPT, complaint rate (track, don't rank on — per the QC Room's own caveat), training completionDaily store checklist · weekly 1:1 with Manju
AnkitaNew-initiative milestones (CPG, Oatmeal Gang — stage-gated, not revenue yet) + CoS cadence healthWeekly cross-team check-ins she already runs

Open questions — genuinely undecided

Needs your call

Is Ankita's double new-initiative load (CPG + Oatmeal Gang) on top of her standing CoS cadence sustainable? That's a real load stacked on a routine role — worth pressure-testing with her directly before this becomes final, not just asserted here.

Needs your call

"Oatmeal Gang" currently means three different things in active use: the internal team WhatsApp ops chat, Rushil's Instagram content bucket, and the proposed gyms/fitness/community program. Worth resolving the name before any program doc gets written on it.

Path to mid-September

  1. This week: lock the org map above — especially the Ankita load question.
  2. Next 1–2 weeks: 30-min working sessions — Rushil first (easiest, agreed), then Abhi, then Manju + Shivaraj + Yiben + Rishabh together as one operating unit — turning the proposed KPIs into something each person actually commits to.
  3. By early September: draft operating-plan docs per person/pod — Oct–Dec detailed, Jan–Mar rough.
  4. Mid-September: review pass before anyone sees theirs.