DISTRICT MANAGER OPERATING SYSTEM
THE ARCHITECTURE
This is not a meeting agenda. It is the complete operating system for
how you run two ghost kitchen locations through two off-site managers.
Six modes. One self-assessment gate. One question bank.
MODES
┌──────────────────────────────────────────────────────────────┐
│ 1. WEEKLY RHYTHM The master calendar. What happens │
│ each day. Time budgets. │
│ │
│ 2. DAILY DATA REVIEW 15-20 min/morning. Dashboards, │
│ metrics, flag/hold decisions. │
│ │
│ 3. LOCATION VISIT Ghost kitchen walkthrough. What to │
│ observe, who to talk to, how long. │
│ │
│ 4. WEEKLY MEETING 1:1 with each AOM. Predict-then- │
│ compare, 5 questions, identity. │
│ │
│ 5. AOM DEVELOPMENT Tier-based coaching. Graduated │
│ exposure. Trust diagnostics. │
│ │
│ 6. CRISIS When stakes exceed the AOM's │
│ current level. Protocol + debrief. │
│ │
│ + SELF-ASSESSMENT GATE Weekly self-grade. Trust audit │
│ monthly. │
│ │
│ + QUESTION BANK Every question you need, organized │
│ by mode and context. │
│ │
└──────────────────────────────────────────────────────────────┘
OPERATIONAL CONTEXT
┌──────────────────────────────────────────────────────────────┐
│ LOCATIONS South Loop (447439) Avondale (273440) │
│ AOMS Lizbeth (South Loop) Angel (Avondale) │
│ ANGEL STATUS Training AOM. Started 2026-05-18. 8-week │
│ graduation gate. Currently IN TRAINING. │
│ Offer: $55k training → $65k post-graduation. │
│ SHIFT LEADS 2-3 per location │
│ TEAM MEMBERS 12-14 per location │
│ HOURS 10 AM - 11 PM (close by midnight/11:30) │
│ │
│ PRIMARY METRICS (company targets from offer letter) │
│ ┌───────────────┬────────────┬───────────────────────┐ │
│ │ Order Errors │ Count │ items missing/wrong │ │
│ │ Error Rate │ 0.75% tgt │ 30-day rolling avg │ │
│ │ Star Ratings │ 4.8+ tgt │ 30-day rolling avg │ │
│ │ Prep Time │ 9 min tgt │ 30-day rolling avg │ │
│ └───────────────┴────────────┴───────────────────────┘ │
│ NOTE: Graduation gate uses the HIGHER of company target │
│ or location's 90-day baseline. If the location is already │
│ below target, the expectation is to maintain the operating │
│ band, not chase the number. Prevent regression first. │
│ │
│ SECONDARY METRICS │
│ ┌───────────────┬────────────┬───────────────────────┐ │
│ │ Attendance │ <20min │ late flags, NCNS │ │
│ │ Labor % │ 17.5% tgt │ of projected sales │ │
│ │ Inventory Var │ <5%/<8% │ dry goods / proteins │ │
│ │ Food Safety │ <41F │ fridge/walk-in temps │ │
│ │ 86 Frequency │ 0 target │ clearest system fail │ │
│ │ Task Comp Rate │ 100% tgt │ per shift │ │
│ └───────────────┴────────────┴───────────────────────┘ │
│ │
│ DATA SOURCES │
│ 7shifts schedule, hours, punch data, labor budget │
│ Platform/POS real-time errors, ratings, prep time │
│ Supabase (2) reflections, execution plans, scorecard │
│ MarketMan inventory, orders, variance │
│ Google Sheet par levels, ADU, order columns (legacy) │
│ Telegram bots inventory, prep, accountability, par, audit │
│ Dashboard training guide: home, analytics, scorecard │
└──────────────────────────────────────────────────────────────┘
AOM SCHEDULING MODEL
REALITY: AOMs are outcome-based. Their offer letter has no
on-site requirement. They are off-site managers by contract.
This means on-site time is a tool you recommend, not a rule
you enforce. The framing matters. "You should be here Monday"
is a mandate. "Monday is when weekend drift surfaces. That is
the highest-leverage day to be on the floor" is a tool.
The AOM decides. You measure results. If the results are good,
their schedule is correct by definition. If results are bad,
on-site time is the first lever you recommend because reports
lie by omission. The floor tells you what the data cannot.
RECOMMENDED ON-SITE TIME (framed as outcome-lever):
THE PRINCIPLE (not the days):
1. One visit after the weekend gap (catch drift).
2. One visit during midweek peak (see the constraint live).
The specific days flex. The principle does not.
Struggling AOM: add a third visit. Not because they are
punished. Because their pattern library is thin. Floor reps
build it faster than data review. Frame it as investment,
not compliance: "You are building the reads that let you
manage from data later. Right now you need to see it."
┌─────────────┬───────────┬────────────────────────────────────┐
│ AOM TIER │ ON-SITE │ HOW TO FRAME IT │
├─────────────┼───────────┼────────────────────────────────────┤
│ STRUGGLING │ 3x/week │ "Your results are not where they │
│ │ │ need to be. The fastest way to fix │
│ │ │ that is floor time. I'd recommend │
│ │ │ three visits this week: one after │
│ │ │ the weekend, one during peak, one │
│ │ │ flex day. Let's see if the numbers │
│ │ │ move." │
│ │ │ │
│ │ │ If they push back: "Your call. The │
│ │ │ results speak. If you can hit [X] │
│ │ │ without being on-site, I don't care │
│ │ │ where you work from." │
│ │ │ │
│ │ │ This is outcome-based management. │
│ │ │ You do not mandate presence. You │
│ │ │ recommend it when outcomes demand │
│ │ │ it. The AOM either accepts the tool │
│ │ │ or demonstrates they can hit the │
│ │ │ numbers without it. Either works. │
├─────────────┼───────────┼────────────────────────────────────┤
│ STABLE │ 2x/week │ Standard recommendation. After │
│ │ │ weekend + midweek peak. Most AOMs │
│ │ │ will naturally settle here because │
│ │ │ they see the value once the pattern │
│ │ │ library builds. │
├─────────────┼───────────┼────────────────────────────────────┤
│ STRONG │ 1-2x/week │ Their choice. Strong AOMs know │
│ │ │ when to be on-site because their │
│ │ │ reads are accurate from data alone. │
│ │ │ They visit when their data tells │
│ │ │ them something needs eyes. That IS │
│ │ │ the proof they are strong. │
└─────────────┴───────────┴────────────────────────────────────┘
EDGE CASE: AOM on-site days don't align to M/W/Th.
This happens. Staffing gaps force the AOM to cover specific
days. A struggling AOM covering Tuesday/Thursday/Saturday
because the SL quit and the schedule is thin.
THE RULE: The days do not matter. The principle does.
1. One visit within 48 hours of the weekend.
2. One visit overlapping a peak window.
3. Third visit (struggling only) on a day they would
otherwise not see the kitchen.
If the AOM is covering shifts, that is a staffing constraint,
not a scheduling preference. Address the constraint (hire,
redistribute) so the AOM can return to the recommended
pattern. An AOM covering shifts is doing shift lead work,
which means the AOM is the bottleneck, which means
development has stalled.
Track how many weeks the AOM is in "shift-covering mode."
If it exceeds 3 weeks, the staffing constraint is now the
binding constraint. Escalate to hire, not to coach.
DM VISIT OVERLAY:
Your visit days serve a different purpose depending on what
you are observing.
TWO TYPES OF DM VISIT:
┌─────────────┬──────────────────────────────────────────────┐
│ TYPE │ PURPOSE │
├─────────────┼──────────────────────────────────────────────┤
│ AOM-PRESENT │ You observe the AOM leading. Their behavior │
│ VISIT │ under real conditions IS the development │
│ │ data. How they walk the floor, what they │
│ │ notice, how they correct, how the team │
│ │ responds to them. This is for struggling │
│ │ and early-stable AOMs. │
├─────────────┼──────────────────────────────────────────────┤
│ AOM-ABSENT │ You observe the system without the AOM. │
│ VISIT │ The kitchen's state when the AOM is not │
│ │ there IS the true measurement of AOM quality. │
│ │ If it runs clean, the system is internalized. │
│ │ If it falls apart, the AOM is the system │
│ │ (presence-dependent, not process-dependent). │
│ │ This is for stable and strong AOMs. │
└─────────────┴──────────────────────────────────────────────┘
┌─────────────┬──────────────────────────────────────────────┐
│ AOM TIER │ DM VISIT TYPE │
├─────────────┼──────────────────────────────────────────────┤
│ STRUGGLING │ 2-3 visits/week. │
│ │ At least one AOM-present (observe them lead). │
│ │ At least one AOM-absent (see the true state). │
│ │ The gap between these two visits tells you │
│ │ whether the AOM is building systems or just │
│ │ being present. │
├─────────────┼──────────────────────────────────────────────┤
│ STABLE │ 1x/week. Prefer AOM-absent visits. │
│ │ The kitchen on an AOM off-day is the real │
│ │ scorecard. If it runs clean, the AOM's │
│ │ systems are working. If not, that is the │
│ │ highest leverage conversation for the meeting.│
├─────────────┼──────────────────────────────────────────────┤
│ STRONG │ 1x/2 weeks. AOM-absent only. │
│ │ Pure verification. The kitchen should not │
│ │ change based on who is present. If it does, │
│ │ the system is not internalized. │
└─────────────┴──────────────────────────────────────────────┘
ON-SHIFT PLAYBOOK (30-second interactions)
When you are on the floor, every interaction is 30 seconds or less.
You are not there to manage. You are there to see, capture, and
occasionally install one thing. Long conversations on the floor
mean you are doing the AOM's job.
THE FIRST 60 SECONDS (every visit):
┌──────────────────────────────────────────────────────────────┐
│ │
│ Walk in. Do not greet anyone yet. Stand where you can see │
│ the whole kitchen. Answer three questions in your head: │
│ │
│ 1. Where is the bottleneck right now? │
│ Which station has the most tickets waiting? │
│ Who is moving slowest? Who looks overwhelmed? │
│ │
│ 2. Is the flow clean or choppy? │
│ Are orders moving left to right (or whatever the │
│ designed flow is) without backtracking? Are people │
│ bumping into each other? Is expo calling back to grill? │
│ │
│ 3. Where is the leader? │
│ Is the shift lead at a station (doing) or walking │
│ (leading)? Is the AOM in the office or on the floor? │
│ The leader's position tells you the system's state. │
│ │
│ That is your read. 60 seconds. Everything else confirms or │
│ contradicts this read. │
│ │
└──────────────────────────────────────────────────────────────┘
SITUATION PLAYBOOK:
What you see → What you do → Why
┌──────────────────────────────────────────────────────────────┐
│ │
│ SITUATION: Orders backing up at one station. │
│ YOU DO: Nothing. Watch what the shift lead does. │
│ CAPTURE: "SL noticed/didn't notice backup at [station] │
│ at [time]. Took [X] seconds to respond." │
│ WHY: The SL's response time IS the system's health. │
│ If they fix it → system works. Capture for praise. │
│ If they don't → capture for weekly meeting P-T-C. │
│ DO NOT: jump in and help the station. That is shift lead │
│ work. Your help teaches the SL to wait for help. │
│ │
├──────────────────────────────────────────────────────────────┤
│ │
│ SITUATION: Team member doing something wrong (wrong build, │
│ wrong portion, wrong temp, cross-contamination). │
│ YOU DO: Look at the AOM or shift lead. Did they see it? │
│ IF THEY SAW IT: Watch how they correct. Capture their │
│ approach for the meeting. Good correction = identity │
│ reflection fuel. Bad correction = coaching fuel. │
│ IF THEY DIDN'T SEE IT: Say nothing. Note it. This is a │
│ measurement gap. The AOM's walkthrough is broken or │
│ their attention is misallocated. P-T-C fuel. │
│ EXCEPTION: Food safety violation. Correct it yourself │
│ immediately. Health code does not wait for coaching. │
│ │
├──────────────────────────────────────────────────────────────┤
│ │
│ SITUATION: Team member approaches you with a problem. │
│ YOU SAY: "Have you talked to [AOM/SL name] about this?" │
│ ALWAYS. Even if you know the answer. Even if it is faster │
│ for you to solve it. Every time you solve a team member's │
│ problem, you subtract authority from the AOM and SL. │
│ The 10 seconds of awkwardness builds the chain of command. │
│ │
├──────────────────────────────────────────────────────────────┤
│ │
│ SITUATION: AOM is in the office during peak. │
│ YOU DO: Note it. Do not go pull them out. │
│ IF STRUGGLING AOM: bring it up at the meeting. │
│ "I visited Wednesday at 12:30. Where were you?" │
│ Let them answer. The question IS the correction. │
│ IF STABLE AOM: could be legitimate (vendor call, emergency │
│ scheduling). Still capture. If pattern repeats 2x, │
│ it becomes a meeting question. │
│ IF STRONG AOM: probably has a reason. Verify the floor ran │
│ fine without them. If yes, that is actually a good sign. │
│ │
├──────────────────────────────────────────────────────────────┤
│ │
│ SITUATION: Prep is behind schedule. │
│ YOU DO: Check the prep list. Is it posted? Is it accurate? │
│ Is the issue a person problem or a system problem? │
│ PERSON: Someone slow, absent, or undertrained. Capture. │
│ SYSTEM: Par levels wrong, delivery late, recipe unclear. │
│ System problems are YOUR problem (or AOM's). Capture │
│ and bring to the weekly meeting as a constraint. │
│ DO NOT: jump on the line and start prepping. The DM │
│ prepping sends the signal that being behind is normal │
│ enough for the DM to accept and compensate. │
│ │
├──────────────────────────────────────────────────────────────┤
│ │
│ SITUATION: Kitchen is clean, running smooth, no issues. │
│ YOU DO: Capture it. "South Loop, Wednesday 12pm. Kitchen │
│ running clean. Expo clear, stations stocked, SL walking │
│ the line. No intervention needed." │
│ This is data too. Consistency is the strongest signal that │
│ the system is working. Use it for identity reflection: │
│ "Three visits in a row, your kitchen ran clean without me │
│ saying a word. That is not luck. That is your standard." │
│ │
├──────────────────────────────────────────────────────────────┤
│ │
│ SITUATION: You see something the AOM should fix, and the │
│ AOM is standing right there not fixing it. │
│ YOU DO (struggling AOM): Narrate. "I'm looking at [X]. │
│ The downstream effect is [Y]. What would you do?" │
│ 30 seconds. Then walk away. Let them process. │
│ YOU DO (stable AOM): Ask. "What are you seeing at [area]?" │
│ If they name it → they know. Trust them to handle it. │
│ If they don't name it → "Look again." Then walk away. │
│ YOU DO (strong AOM): Say nothing. If they don't catch it │
│ today, it becomes a P-T-C situation for the meeting. │
│ Strong AOMs get the dignity of handling their own floor. │
│ │
├──────────────────────────────────────────────────────────────┤
│ │
│ SITUATION: Shift change happening (team arriving/leaving). │
│ YOU DO: Watch the handoff. Is there a handoff? Does the │
│ outgoing SL brief the incoming SL? Are open items passed? │
│ A bad handoff = every problem restarts from zero next shift.│
│ CAPTURE: "Shift handoff at [time]. [Happened/didn't]. │
│ SL briefed: [yes with detail / yes superficial / no]." │
│ This is one of the highest leverage observations. Most │
│ recurring problems trace back to broken handoffs. │
│ │
├──────────────────────────────────────────────────────────────┤
│ │
│ SITUATION: Disagreement between team members on the floor. │
│ YOU DO: Watch the AOM or SL handle it. Do not intervene. │
│ Even if they handle it badly. The only exception is if it │
│ escalates to yelling or physical aggression. Short of that, │
│ the AOM learning to handle conflict IS the development. │
│ Capture their approach. Debrief at meeting if needed. │
│ │
└──────────────────────────────────────────────────────────────┘
THE RULE: If your interaction takes more than 30 seconds, you
are managing, not leading. Capture it. Bring it to the meeting.
The floor is for observation. The meeting is for development.
The only exception is a safety or crisis event.
AOM OPERATING SYSTEM (reference)
The AOM has their own operating system: _strategy/aom-operating-system.md
That document is the AOM's perspective. It covers:
1. On-site operations (measure, compare+correct, coach one thing)
2. Off-site management (data, scheduling, ordering, admin)
3. Weekly meeting prep (how to show up prepared for Q1-Q5)
4. Shift lead development (cascading the coaching system down)
The DM OS you are reading is YOUR operating system.
The AOM OS is THEIR operating system.
Same system, two perspectives.
When evaluating an AOM, check their execution against their OS.
When coaching an AOM, reference their OS, not yours.
OFF-SHIFT ANALYSIS GUIDE (what each number means)
When you are off-site, you lead through data. But data is only
useful if you know what it means and what to do about it.
This is the translation layer. Each metric has a signal, a
threshold, and an action. 30 seconds per metric.
PRIMARY METRICS ANALYSIS:
┌──────────────────────────────────────────────────────────────┐
│ │
│ ORDER ERRORS (count) │
│ │
│ What it measures: items missing, wrong, or remade. │
│ Where to find it: Platform/POS dashboard, daily report. │
│ │
│ READING THE NUMBER: │
│ 0-2 errors/day → Clean. System working. │
│ 3-5 errors/day → Drift. Something changed this week. │
│ Check: new team member? menu change? │
│ recipe clarity? station layout shift? │
│ 6+ errors/day → Broken. A process is failing. │
│ Check: is it one station or all? │
│ One station = person or recipe issue. │
│ All stations = systemic (rush, staff). │
│ │
│ PATTERN: One-day spike = incident (hold). │
│ 2-day trend = emerging (investigate). │
│ 3+ days = systemic (act or meeting). │
│ │
│ ACTION: Hold for meeting UNLESS 6+ errors two days in a │
│ row. Then ask AOM one question: "What changed at expo?" │
│ │
├──────────────────────────────────────────────────────────────┤
│ │
│ ERROR RATE (%) │
│ │
│ What it measures: errors / total orders. Normalizes volume. │
│ Target: below 0.9%. │
│ │
│ WHY THIS MATTERS MORE THAN COUNT: │
│ 10 errors on 200 orders (5%) is a disaster. │
│ 10 errors on 1,500 orders (0.67%) is acceptable. │
│ Always check rate, not just count. │
│ │
│ READING THE NUMBER: │
│ < 0.7% → Excellent. Reinforce at meeting. │
│ 0.7% - 0.9% → Target zone. Hold. │
│ 0.9% - 1.5% → Slipping. Capture for meeting. │
│ Ask: "What's the error TYPE?" │
│ Missing items = prep/par issue. │
│ Wrong items = recipe/training issue. │
│ Late items = flow/bottleneck issue. │
│ > 1.5% → Broken. Same-day contact with AOM. │
│ "Error rate hit [X]%. Walk me through │
│ what happened during peak." │
│ │
├──────────────────────────────────────────────────────────────┤
│ │
│ STAR RATINGS (daily average) │
│ Target: 4.8+ │
│ │
│ READING THE NUMBER: │
│ 4.8+ → Healthy. No action. │
│ 4.5 - 4.8 → Watch. Check if it is one platform │
│ dragging the average or all platforms. │
│ Read the 1-2 star reviews. What do │
│ they name? That is the constraint. │
│ < 4.5 → Reputation risk. Same-day ask: │
│ "Pull up today's low reviews. What do │
│ they say?" The review text is the │
│ diagnostic. Not the number. │
│ │
│ TRAP: Ratings lag. Today's rating reflects food sent 24-72 │
│ hours ago. If error rate spiked Thursday, expect ratings │
│ drop Saturday-Monday. Do not panic on a single-day drop. │
│ Three-day trends are the signal. │
│ │
├──────────────────────────────────────────────────────────────┤
│ │
│ PREP TIME (minutes, peak-weighted) │
│ │
│ READING THE NUMBER: │
│ At or below target → No action. │
│ 5-10% above target → Capacity pressure. Check: did volume │
│ spike or did throughput drop? │
│ Volume spike = good problem (demand). │
│ Throughput drop = bad problem (drift). │
│ 20%+ above target → Flow is broken. Something upstream │
│ failed. Common causes: │
│ - Prep not done before peak. │
│ - Station layout changed or supplies │
│ not restocked. │
│ - Short staffed at a critical station. │
│ Bring to meeting with specific time │
│ window: "Prep time spiked between │
│ 11:30 and 1pm. What happened?" │
│ │
└──────────────────────────────────────────────────────────────┘
SECONDARY METRICS ANALYSIS:
┌──────────────────────────────────────────────────────────────┐
│ │
│ ATTENDANCE │
│ What to look for: late arrivals (>20 min), NCNS (no call │
│ no show), early clock-outs. │
│ Signal: Same person late 2x in a week = conversation │
│ needed (AOM's job, not yours). NCNS = immediate staffing │
│ gap. If AOM handled it without escalating = good signal. │
│ If AOM did not handle it at all = P-T-C fuel. │
│ │
│ LABOR % │
│ Target: 17.5% of projected sales. │
│ Above 19% → Overstaffed or low sales. Check both. │
│ Below 15% → Understaffed. Team burning out. Error rate │
│ will spike in 1-2 weeks (fatigue lag). │
│ The number alone is not enough. Cross-reference with │
│ order volume: high labor % on a slow day = scheduling │
│ problem. High labor % on a busy day = pricing problem. │
│ │
│ INVENTORY VARIANCE │
│ Dry goods target: < 5%. Proteins target: < 8%. │
│ Above threshold → Three possible causes: │
│ 1. Overportioning (most common). Check high-cost items. │
│ 2. Waste (expired, dropped, burned). Check waste log. │
│ 3. Theft (rare but real). Variance on expensive items │
│ only, no waste log entries. Sensitive conversation. │
│ Do not accuse. Bring the number to the meeting: │
│ "Chicken breast variance was 12% this week. Walk me │
│ through what could cause that." │
│ │
│ 86 FREQUENCY (stockouts) │
│ Target: 0. │
│ Any 86 = a system failure, not a one-time miss. │
│ Trace backward: Was it an ordering issue (par too low)? │
│ Delivery issue (vendor short)? Usage issue (overportioning │
│ or unexpected volume)? Each cause has a different fix. │
│ 86 events are the clearest test of whether the AOM's │
│ ordering and par system actually works. │
│ │
│ TASK COMPLETION RATE │
│ Target: 100% per shift. │
│ Below 90% → Which tasks are being skipped? If it is │
│ always the same tasks, those tasks may be wrong (too many, │
│ unclear, or placed at the wrong time in the shift). │
│ If different tasks each time → accountability issue. The │
│ SL is not verifying. That is a shift lead coaching problem, │
│ which is the AOM's problem, which becomes your meeting Q. │
│ │
└──────────────────────────────────────────────────────────────┘
CROSS-LOCATION COMPARISON (weekly):
┌──────────────────────────────────────────────────────────────┐
│ │
│ Pull the same metrics for both locations side by side. │
│ The comparison reveals what a single location cannot: │
│ │
│ SAME PROBLEM BOTH LOCATIONS: │
│ → Systemic. Not the AOM. Check: menu change? vendor issue? │
│ platform policy? Supply chain? If both locations show │
│ the same spike, the cause is above the AOM level. │
│ │
│ PROBLEM AT ONE, NOT THE OTHER: │
│ → AOM-specific or location-specific. Check: staffing diff? │
│ Volume diff? If similar volume + staffing but different │
│ results, the AOM's process is the variable. │
│ This is the highest leverage observation for the meeting. │
│ "South Loop error rate was 0.6% this week. Avondale was │
│ 1.3%. Same menu, similar volume. What's different?" │
│ │
│ ONE SOLVING WHAT THE OTHER STRUGGLES WITH: │
│ → Transfer event. Trigger the cross-location protocol from │
│ Mode 1 Friday close. This is how the DM creates value │
│ that neither AOM can create alone. │
│ │
└──────────────────────────────────────────────────────────────┘
OFF-SHIFT COMMUNICATION RULES:
┌──────────────────────────────────────────────────────────────┐
│ │
│ DEFAULT: Do not text the AOM about data. │
│ Hold it for the meeting. Your restraint builds their │
│ independence. Every text is a signal that they cannot │
│ be trusted to find it themselves. │
│ │
│ TEXT ONLY WHEN: │
│ - Star rating drops below 4.5 (reputation damage active) │
│ - Error rate doubles overnight (system broken, not drift) │
│ - 86 event detected (supply chain issue, time-sensitive) │
│ - Food safety violation visible in data (temp logs) │
│ - NCNS causing same-day staffing gap │
│ │
│ TEXT FORMAT (one question, not a lecture): │
│ "Error rate hit 2.1% yesterday. What happened during peak?" │
│ NOT: "Error rate was 2.1%. You need to check your expo │
│ process and make sure the team is following the build │
│ cards. Also review the closing checklist because..." │
│ │
│ One question. Let them diagnose. Their diagnosis tells │
│ you more than your lecture ever would. │
│ │
└──────────────────────────────────────────────────────────────┘
UPSTREAM CONSTRAINTS (what actually builds elite managers)
Six constraints. Every protocol in this OS trains at least one.
If a protocol does not map to a constraint, it is waste. Cut it.
CONSTRAINT MECHANISM TRAINED BY
1. Pattern Caudate nucleus tracks MODE 4 predict-then-compare.
Recognition prediction error. Commitment MODE 3 captured situations.
Under Stakes before reveal activates error 8x encoding vs reflective
correction. The gap IS the rep. learning.
2. Self-Regulation Regulation builds through MODE 5 graduated exposure.
Under Threat surviving stress, not being MODE 6 crisis debrief.
told "you can handle it." Each survived decision
Each override shrinks it. expands the window.
3. Standard If they solve but cannot teach, MODE 4 monthly external-
Externalization they are a permanent bottleneck. ization check. "Could a
Portability is the test. new hire use this without
calling you?"
4. Control Override activates ACC conflict. MODE 4 the 5 questions.
Paradox Repeated overrides: dopamine MODE 2 data flag → hold.
drops, ToM disengages, leader MODE 3 redirect to AOM.
degrades to doer. Cumulative. Every restraint is a rep.
5. Mirror Mirror neurons copy action, not MODE 3 on-site behavior.
Circuit language. ~15% signal loss per MODE 6 crisis regulation.
layer. Your behavior under How you handle bad news
stress propagates with highest programs their crisis
fidelity. response permanently.
6. Trust Three independent computations. MODE 3 consistent behavior.
Layers Missing any one breaks the MODE 4 narrate reasoning.
whole system differently. MODE 2 data-driven calls.
Built daily, broken once.
┌─────────────┬──────────────────────────────┐
│ COMPETENCE │ Built: right on big calls. │
│ │ Broken: wrong on big calls. │
├─────────────┼──────────────────────────────┤
│ CHARACTER │ Built: group-serving acts. │
│ │ Broken: ONE self-serving act.│
├─────────────┼──────────────────────────────┤
│ JUDGMENT │ Built: transparent reasoning │
│ │ that succeeds. Broken: │
│ │ opaque decisions that fail. │
└─────────────┴──────────────────────────────┘
MODE 1: WEEKLY RHYTHM
The master calendar. Every other mode plugs into this.
┌─────────┬──────────────────────────────────────────────────────────┐
│ DAY │ WHAT HAPPENS │
├─────────┼──────────────────────────────────────────────────────────┤
│ │ │
│ EVERY │ DAILY DATA REVIEW (Mode 2) 15-20 min │
│ DAY │ Morning. Before any other work. │
│ │ Check dashboards. Flag or hold. │
│ │ │
├─────────┼──────────────────────────────────────────────────────────┤
│ │ │
│ MON │ WEEKLY LAUNCH 30 min │
│ │ │
│ │ 1. Pull analytics for both locations (Mode 2 deep) │
│ │ 2. Review last week's execution plan targets │
│ │ - Did the metric move? By how much? │
│ │ - Was the constraint correctly identified? │
│ │ 3. Set 1-2 specific targets for this week per location │
│ │ - Each target must be metric-linked and time-bound │
│ │ 4. Identify which location gets priority visits │
│ │ this week (performance tier drives this) │
│ │ 5. Capture 1-2 situations from data for predict-then- │
│ │ compare in this week's meetings │
│ │ │
│ │ OUTPUT: week's priorities written down. │
│ │ Two targets per location. Visit schedule set. │
│ │ │
├─────────┼──────────────────────────────────────────────────────────┤
│ │ │
│ TUE │ LOCATION VISIT: PRIORITY LOCATION 1-2 hrs │
│ or │ (Mode 3) │
│ WED │ │
│ │ Walk in with ONE data question from Monday's review. │
│ │ Not "how's it going." A specific question: │
│ │ "Error rate spiked Thursday. What happened at expo?" │
│ │ "Prep time is trending up. Walk me through prep flow." │
│ │ │
│ │ Ghost kitchen visit protocol (see Mode 3). │
│ │ │
├─────────┼──────────────────────────────────────────────────────────┤
│ │ │
│ WED │ MIDWEEK DATA CHECK 20 min │
│ or │ │
│ THU │ Deeper than daily review. Focused on this week's │
│ │ two targets per location: │
│ │ - Are the targets tracking? │
│ │ - If not: effort issue or system issue? │
│ │ │
│ │ DECISION TREE: │
│ │ ┌─────────────────────┐ │
│ │ │ Target tracking? │ │
│ │ ├──── YES ───────────→│ Hold. Bring to meeting as win.│
│ │ ├──── NO ────────────→│ Effort or system? │
│ │ │ ├── EFFORT ─────→│ Capture for meeting. Do NOT │
│ │ │ │ │ text the AOM about it. │
│ │ │ └── SYSTEM ─────→│ Investigate. Is a process │
│ │ │ │ broken? Tool missing? Begin │
│ │ │ │ fix if you can. Bring to │
│ │ │ │ meeting if it needs AOM. │
│ │ └──────────────────────┘ │
│ │ │
├─────────┼──────────────────────────────────────────────────────────┤
│ │ │
│ THU │ WEEKLY MEETING: AOM 1 15-20 min │
│ or │ (Mode 4) │
│ FRI │ │
│ │ WEEKLY MEETING: AOM 2 15-20 min │
│ │ │
│ │ Both meetings happen this day or split across two. │
│ │ Same day, same time, every week. No rescheduling. │
│ │ │
├─────────┼──────────────────────────────────────────────────────────┤
│ │ │
│ FRI │ WEEKLY CLOSE 15 min │
│ │ │
│ │ 1. Self-assessment gate (Mode 7) │
│ │ 2. Cross-location pattern check: │
│ │ - Same error type failing at both locations? │
│ │ - One location solving what the other struggles │
│ │ with? │
│ │ IF YES: trigger same-week transfer protocol: │
│ │ a. Ask solving AOM to write it down: │
│ │ "What did you change? What would you tell │
│ │ [other AOM]?" │
│ │ b. Gate: "Could they use this without you?" │
│ │ YES → pass to other AOM next meeting. │
│ │ NO → refine until portable. │
│ │ c. Other AOM implements with support for 1-2 │
│ │ weeks, then independently. │
│ │ Do NOT wait for the monthly externalization │
│ │ check. Cross-location transfer happens the same │
│ │ week the pattern is detected. │
│ │ 3. Escalation budget check (see Mode 5 tracking) │
│ │ 4. Update execution plans for next week │
│ │ 5. Capture any remaining situations for next meeting │
│ │ │
│ │ OPTIONAL: second visit to struggling location if │
│ │ midweek data flagged something that needs eyes. │
│ │ │
├─────────┼──────────────────────────────────────────────────────────┤
│ │ │
│ SAT/ │ MINIMAL. Data review only. 10 min │
│ SUN │ No visits unless crisis. │
│ │ No texts unless escalation budget exceeded. │
│ │ │
│ │ Your absence on weekends IS data. │
│ │ If the kitchen runs clean without you = system works. │
│ │ If it falls apart = self-sustaining cycle not built. │
│ │ Either way, the answer comes from Monday's data, not │
│ │ from a weekend visit. │
│ │ │
└─────────┴──────────────────────────────────────────────────────────┘
WEEKLY TIME BUDGET
┌────────────────────────────────────┬──────────┐
│ Daily data review (5x weekday) │ 75-100m │
│ Monday weekly launch │ 30m │
│ Location visit (1-3x) │ 60-180m │
│ Weekly meetings (2x AOM) │ 30-40m │
│ Midweek data check │ 20m │
│ Friday weekly close │ 15m │
│ Weekend data review (2x) │ 20m │
│ Situation capture (ongoing) │ 10m │
├────────────────────────────────────┼──────────┤
│ WEEKLY TOTAL │ 4-6 hrs │
└────────────────────────────────────┴──────────┘
That is the full investment. 4-6 hours per week to run two locations
through two managers. The rest of the time, the system runs itself.
If you are spending more than 6 hours, you are overriding, not leading.
VISIT FREQUENCY (performance-based):
┌─────────────┬──────────┬────────────────────────────────────┐
│ TIER │ VISITS │ SIGNAL │
├─────────────┼──────────┼────────────────────────────────────┤
│ STRUGGLING │ 2-3x/wk │ Primary metrics below target for │
│ │ │ 2+ consecutive weeks. AOM asking │
│ │ │ permission instead of reporting. │
│ │ │ Escalation budget exceeded. │
├─────────────┼──────────┼────────────────────────────────────┤
│ STABLE │ 1x/wk │ Primary metrics near target. │
│ │ │ AOM reporting decisions. Occasional │
│ │ │ misses but self-correcting. │
├─────────────┼──────────┼────────────────────────────────────┤
│ STRONG │ 1x/2wk │ Primary metrics consistently at │
│ │ │ or above target. AOM running full │
│ │ │ negentropy cycle independently. │
│ │ │ Spot audits confirm standards hold. │
└─────────────┴──────────┴────────────────────────────────────┘
TIER TRANSITION RULES:
- Struggling → Stable: 2 consecutive weeks with all primary
metrics at or above target AND escalation budget respected.
- Stable → Strong: 4 consecutive weeks at target AND AOM
independently identifies and fixes a constraint you did not
flag. (Proves the model is internalized, not just compliant.)
- Strong → Stable: any primary metric below target for 2 weeks.
- Stable → Struggling: any primary metric below target for 3
weeks OR escalation budget exceeded 2x in a month.
- Strong → Struggling: skip Stable only if a crisis was handled
poorly (system failed, not just a bad week).
MODE 2: DAILY DATA REVIEW
15-20 minutes. Every weekday morning. Before anything else.
This is the MEASURE operation of your negentropic cycle as DM.
Without this, you are leading blind. Every downstream decision
depends on this input.
PURPOSE: Know the state of both locations before they tell you.
When the AOM texts you a problem, you already know.
When the AOM does not text you a problem that exists,
you know that too.
SEQUENCE:
┌──────────────────────────────────────────────────────────────┐
│ │
│ STEP 1: PRIMARY METRICS (5 min) │
│ │
│ Check yesterday's numbers for both locations: │
│ │
│ ┌───────────────┬──────────┬──────────┬──────────┐ │
│ │ METRIC │ S. LOOP │ AVONDALE │ FLAG? │ │
│ ├───────────────┼──────────┼──────────┼──────────┤ │
│ │ Order Errors │ │ │ >target │ │
│ │ Error Rate │ │ │ >0.9% │ │
│ │ Star Rating │ │ │ <4.8 │ │
│ │ Prep Time │ │ │ >target │ │
│ └───────────────┴──────────┴──────────┴──────────┘ │
│ │
│ SOURCE: Platform/POS dashboards. │
│ │
│ NO FLAGS → move to step 2. │
│ FLAGS → note which metric, which location. Do NOT text │
│ the AOM. This goes to the weekly meeting or the midweek │
│ check, depending on severity. │
│ │
│ SEVERITY OVERRIDE (act now, do not hold for meeting): │
│ - Star rating drops below 4.5 (reputation damage) │
│ - Error rate doubles overnight (system broken, not drift) │
│ - 86 event (stockout = immediate supply chain issue) │
│ - Food safety temp out of range (health code risk) │
│ │
├──────────────────────────────────────────────────────────────┤
│ │
│ STEP 2: SECONDARY METRICS (5 min) │
│ │
│ Quick scan. Not every day for all of these. Rotate focus. │
│ │
│ MON: Attendance (weekend late flags, NCNS from Fri-Sun) │
│ TUE: Labor % (are we at 17.5% target? OT building?) │
│ WED: Inventory variance (any items >5%/8% threshold?) │
│ THU: Task completion rates (which shifts incomplete?) │
│ FRI: 86 frequency this week (any stockouts?) │
│ │
│ SOURCE: 7shifts (attendance, labor), MarketMan (inventory), │
│ Dashboard (tasks, accountability), Telegram bots (par, 86). │
│ │
├──────────────────────────────────────────────────────────────┤
│ │
│ STEP 3: SIGNALS (5 min) │
│ │
│ Quick scan of signal channels: │
│ │
│ - Dashboard home page: reflections pending? accountability │
│ open items? training episodes in progress? │
│ - Telegram bot channels: any alerts overnight? │
│ - AOM messages: any escalations? (count against budget) │
│ │
│ DECISION: For each signal, classify: │
│ │
│ ┌────────────────┬─────────────────────────────────┐ │
│ │ HOLD │ Bring to weekly meeting. │ │
│ │ │ AOM does not know you noticed. │ │
│ │ │ Becomes predict-then-compare fuel.│ │
│ ├────────────────┼─────────────────────────────────┤ │
│ │ INVESTIGATE │ You need more data before │ │
│ │ │ deciding. Check a second source. │ │
│ │ │ Still do not text the AOM. │ │
│ ├────────────────┼─────────────────────────────────┤ │
│ │ ACT NOW │ Severity override triggered. │ │
│ │ │ Contact AOM. One question: │ │
│ │ │ "What happened with [X]?" │ │
│ │ │ Let them explain. Then decide. │ │
│ └────────────────┴─────────────────────────────────┘ │
│ │
│ DEFAULT IS HOLD. Not act. Your restraint is the test. │
│ Every unnecessary text is a control paradox violation. │
│ │
│ CAPTURE: If you notice something worth discussing, │
│ write a 30-second phone note: │
│ "[Location]. [Day]. [What the data showed]. [Why it │
│ matters]. [What I would have done]." │
│ This becomes a predict-then-compare rep in the meeting. │
│ │
└──────────────────────────────────────────────────────────────┘
MODE 3: LOCATION VISIT
Ghost kitchen walkthrough protocol. Not a casual check-in.
Every visit answers a specific question identified from data.
PRE-VISIT PREP (5 min, before you leave):
┌──────────────────────────────────────────────────────────────┐
│ │
│ 1. What is the ONE question I am here to answer? │
│ Must come from daily data review or midweek check. │
│ Examples: │
│ - "Error rate spiked Thursday. What's happening at expo?"│
│ - "Prep time trending up. Is it a staffing gap or │
│ process issue?" │
│ - "Lizbeth avoided the conversation with [team member]. │
│ Where is her regulation window?" │
│ - No question = no visit. A visit without purpose is │
│ surveillance, not leadership. │
│ │
│ 2. Review this location's primary metrics (30 sec). │
│ Know the numbers before you walk in. │
│ The AOM should not be the one telling you the score. │
│ │
│ 3. Review AOM's current development tier. │
│ This determines your behavior on the floor (see below). │
│ │
└──────────────────────────────────────────────────────────────┘
ON-SITE PROTOCOL:
┌──────────────────────────────────────────────────────────────┐
│ │
│ PHASE 1: ARRIVE + OBSERVE (10-15 min) │
│ │
│ Do not announce your arrival with a meeting. │
│ Walk the floor first. Observe before talking. │
│ │
│ GHOST KITCHEN WALKTHROUGH CHECKLIST: │
│ │
│ ┌─ STATIONS ─────────────────────────────────────────┐ │
│ │ □ Each station staffed and set up correctly? │ │
│ │ □ Mise en place complete? Containers filled? │ │
│ │ □ Station flow: does the layout match the volume? │ │
│ │ □ Cross-contamination risks? (raw/cooked zones) │ │
│ └────────────────────────────────────────────────────┘ │
│ │
│ ┌─ PREP ─────────────────────────────────────────────┐ │
│ │ □ Prep list posted and being followed? │ │
│ │ □ Par levels matching what was ordered? │ │
│ │ □ Any items close to 86? (supply chain signal) │ │
│ │ □ Prep timing: on schedule or behind? │ │
│ └────────────────────────────────────────────────────┘ │
│ │
│ ┌─ FOOD SAFETY ──────────────────────────────────────┐ │
│ │ □ Fridge/walk-in temps (<41F)? │ │
│ │ □ Hot holder temps at spec? │ │
│ │ □ Date labels current? FIFO followed? │ │
│ │ □ Handwashing compliance? Glove changes? │ │
│ └────────────────────────────────────────────────────┘ │
│ │
│ ┌─ FLOW ─────────────────────────────────────────────┐ │
│ │ □ Orders flowing without bottleneck? │ │
│ │ □ Where is the constraint right now? │ │
│ │ (Which station/person is the throughput limit?) │ │
│ │ □ Any obvious waste? (overportioning, remakes) │ │
│ │ □ Team communication: are callouts happening? │ │
│ └────────────────────────────────────────────────────┘ │
│ │
│ ┌─ PEOPLE ───────────────────────────────────────────┐ │
│ │ □ Where is the AOM? (on floor or in office?) │ │
│ │ □ Where is the shift lead? (walking or stationed?)│ │
│ │ □ Team energy: engaged or going through motions? │ │
│ │ □ Anyone struggling? (speed, accuracy, confidence) │ │
│ └────────────────────────────────────────────────────┘ │
│ │
│ CAPTURE everything you notice. 30-second phone notes. │
│ These become predict-then-compare fuel. │
│ │
├──────────────────────────────────────────────────────────────┤
│ │
│ PHASE 2: WALK WITH AOM (15-30 min) │
│ │
│ Ask the AOM to walk you through their floor. │
│ Do NOT impose your route. Their walkthrough reveals: │
│ - What they measure (where their attention goes) │
│ - What they miss (where their attention does not go) │
│ - How they narrate (depth of their mental model) │
│ │
│ BEHAVIOR BY AOM TIER: │
│ │
│ ┌─ STRUGGLING AOM ───────────────────────────────────┐ │
│ │ YOU: Narrate your reasoning aloud. │ │
│ │ "I'm looking at [X] because [Y] downstream." │ │
│ │ "The constraint right now is [station/person] │ │
│ │ because [reason]." │ │
│ │ │ │
│ │ PURPOSE: Install the mental model. They learn │ │
│ │ HOW you see, not just WHAT you see. │ │
│ │ This is JUDGMENT TRUST building. │ │
│ │ │ │
│ │ DO NOT: solve problems for them. Narrate what │ │
│ │ you observe, then ask "What would you do?" │ │
│ └─────────────────────────────────────────────────────┘ │
│ │
│ ┌─ STABLE AOM ───────────────────────────────────────┐ │
│ │ YOU: Ask questions. Minimal narration. │ │
│ │ "What are you seeing at expo?" │ │
│ │ "Where's your constraint today?" │ │
│ │ "What would you change about this setup?" │ │
│ │ │ │
│ │ PURPOSE: Test their model. Let their computation │ │
│ │ run. Compare their reads to yours silently. │ │
│ │ The gap between your read and theirs is data │ │
│ │ for the weekly meeting. │ │
│ │ │ │
│ │ DO NOT: narrate unless they are clearly stuck. │ │
│ │ Your silence IS the graduated exposure. │ │
│ └─────────────────────────────────────────────────────┘ │
│ │
│ ┌─ STRONG AOM ───────────────────────────────────────┐ │
│ │ YOU: Observe. Ask almost nothing. │ │
│ │ Let them run the floor. Watch their team. │ │
│ │ Watch how the team responds to them. │ │
│ │ │ │
│ │ PURPOSE: Verify the system is self-sustaining. │ │
│ │ If the floor runs clean with you watching and │ │
│ │ not speaking, the negentropy cycle is internal. │ │
│ │ │ │
│ │ Your silence says: "I trust this." That is the │ │
│ │ highest signal you can send. │ │
│ │ │ │
│ │ CAPTURE: Note what they do well. This is fuel │ │
│ │ for identity reflection in the meeting. │ │
│ └─────────────────────────────────────────────────────┘ │
│ │
│ REDIRECT RULE (all tiers): │
│ If a team member brings a problem to you: │
│ "Have you talked to [AOM name] about this?" │
│ Never solve for the team when the AOM is present. │
│ Every bypass you allow degrades the AOM's authority. │
│ │
├──────────────────────────────────────────────────────────────┤
│ │
│ PHASE 3: ANSWER YOUR QUESTION (5 min) │
│ │
│ Before you leave, answer the question you came with: │
│ │
│ "Error rate spiked Thursday. What's happening at expo?" │
│ → Answer: [what you found] │
│ → Binding constraint: [what's actually causing it] │
│ → Action: [hold for meeting / investigate more / act now] │
│ │
│ If you cannot answer the question from this visit, you │
│ need a different data source, not a longer visit. │
│ │
├──────────────────────────────────────────────────────────────┤
│ │
│ POST-VISIT CAPTURE (2 min, in the car) │
│ │
│ Phone note, 30 seconds per situation: │
│ "[Location]. [Date]. [What I observed]. [What I would │
│ have done]. [Why]. [What the AOM did or did not notice]." │
│ │
│ These become predict-then-compare reps in the next meeting. │
│ Target: 2-3 captured situations per visit. │
│ │
└──────────────────────────────────────────────────────────────┘
VISIT DURATION BY TIER:
┌─────────────┬──────────┬──────────────────────────────────────┐
│ TIER │ TIME │ WHAT YOU DO │
├─────────────┼──────────┼──────────────────────────────────────┤
│ STRUGGLING │ 1.5-2 hr │ Full walkthrough. Narrate reasoning. │
│ │ │ Work alongside on one problem. │
│ │ │ Capture 3+ situations. │
├─────────────┼──────────┼──────────────────────────────────────┤
│ STABLE │ 45-60min │ AOM-led walkthrough. Questions only. │
│ │ │ Spot check one metric. Capture 2. │
├─────────────┼──────────┼──────────────────────────────────────┤
│ STRONG │ 30 min │ Observe. Minimal interaction. │
│ │ │ Verify self-sustaining. Capture 1. │
└─────────────┴──────────┴──────────────────────────────────────┘
MODE 4: WEEKLY MEETING
DURATION 15-20 MINUTES PER AOM
CADENCE Weekly (same day, same time, no rescheduling)
LOCATION Private. Not on the floor. Not in earshot of team.
Phone/video is fine for ghost kitchen DM.
┌──────────────────────────────────────────────────────────┐
│ │
│ PHASE 1: PREDICT-THEN-COMPARE (5 min) │
│ │
│ MECHANISM: Caudate nucleus tracks prediction error. │
│ Commitment before reveal = 8x encoding quality. │
│ This is not a quiz. It is the neurological mechanism │
│ by which the pattern library grows. │
│ │
│ STEP 1: DIAGNOSE LAST WEEK │
│ │
│ "You said you would try [X] to fix [Y]. You can see │
│ the numbers. Did it work? Why or why not?" │
│ │
│ They commit to a CAUSE for why the metric moved or │
│ did not move. Then reveal what you found from your │
│ captures or data review. │
│ │
│ The gap between their diagnosis and the real cause │
│ is where the pattern library grows. │
│ │
│ STEP 2: NEW SITUATIONS (from your captures) │
│ │
│ Present 2-3 captured situations from the past week. │
│ Sources: your visit observations, daily data flags, │
│ midweek check findings, other location's problems. │
│ │
│ EXECUTION PROTOCOL (8 steps, do not skip any): │
│ │
│ 1. Present the situation. No hints. No context that │
│ reveals the answer. Just: "This happened at │
│ [location] on [day]. [Description]." │
│ │
│ 2. "What would YOU do?" Force commitment. │
│ Do not accept "it depends." Make them pick. │
│ │
│ 3. Wait 5 seconds for thinking. Silence is reps. │
│ The urge to fill silence is YOUR control paradox. │
│ │
│ 4. Listen to their full answer. Do not interrupt. │
│ Note their answer mentally for trend tracking. │
│ │
│ 5. "Here's what happened / what I would have done." │
│ Reveal. Factual. No judgment tone. │
│ │
│ 6. 10 seconds silence. Let the gap land. │
│ Do NOT lecture. Do NOT explain why they were wrong. │
│ The silence is where the encoding happens. │
│ │
│ 7. "What's the delta?" AOM articulates the gap │
│ themselves. Their narration of the gap encodes │
│ deeper than your explanation of it. │
│ │
│ 8. Move to next situation or Phase 2. │
│ │
│ SLOPPY EXECUTION KILLS THE MULTIPLIER: │
│ - Hinting before commit = no prediction error (2x, not 8x)│
│ - Not waiting for commitment = passive learning (1x) │
│ - Lecturing after reveal = your encoding, not theirs │
│ - Skipping "what's the delta" = gap unprocessed │
│ │
│ GAP TRACKING: │
│ ┌──────────────┬──────────────┬──────────────┐ │
│ │ WEEK │ GAP SIZE │ SIGNAL │ │
│ ├──────────────┼──────────────┼──────────────┤ │
│ │ 1-2 │ 60%+ │ Expected │ │
│ │ 3-4 │ 40% │ Installing │ │
│ │ 5-6 │ 30% │ Building │ │
│ │ 7-8 │ 20% │ Converging │ │
│ │ 9-12 │ <20% │ Shared model│ │
│ │ 13+ │ <10% │ Autonomous │ │
│ └──────────────┴──────────────┴──────────────┘ │
│ │
├──────────────────────────────────────────────────────────┤
│ │
│ PHASE 2: THE 5 QUESTIONS (5-8 min) │
│ │
│ Q1: "Scan these categories. What did you notice │
│ this week?" │
│ │
│ GUIDED CATEGORIES: │
│ ┌─────────────┬─────────────┬─────────────┐ │
│ │ ORDER ERRORS │ PREP TIME │ STAR RATINGS│ │
│ ├─────────────┼─────────────┼─────────────┤ │
│ │ FOOD SAFETY │ ATTENDANCE │ CONDUCT │ │
│ └─────────────┴─────────────┴─────────────┘ │
│ Top row = output metrics. │
│ Bottom row = input causes. │
│ │
│ If blank on all six → MEASURE operation failed. │
│ Their measurement routine is broken. Rebuild it │
│ before proceeding. Nothing downstream works │
│ without this. │
│ │
│ Q2: "What's the biggest problem?" │
│ The one thing that moved (or failed to move) a │
│ metric the most this week. Not what frustrated │
│ them. Not what's freshest in memory. The data. │
│ │
│ If they list many → "If you could only fix one." │
│ If personal vs metric-driven → redirect to data. │
│ │
│ Q3: "What's the constraint?" │
│ The upstream cause. Not the symptom. │
│ │
│ BINDING TEST: "If we fixed that tomorrow, │
│ would the metric move?" │
│ YES = binding (real constraint). NO = slack (noise).│
│ │
│ If symptoms → redirect ONCE: "Upstream. What │
│ causes that?" │
│ If still symptoms → they need more shared input │
│ (narrate reasoning next visit). │
│ │
│ Q4: "What are you going to try?" │
│ If stuck → "What's one small thing you could do │
│ today?" │
│ Their answer does not need to be your answer. │
│ A worse plan they own > a better plan you gave. │
│ │
│ Q5: "When are you doing it?" │
│ "This week" is NOT an answer. │
│ Must produce: day, time, duration. │
│ If they cannot specify → shrink scope. │
│ │
├──────────────────────────────────────────────────────────┤
│ │
│ PHASE 3: IDENTITY REFLECTION (2 min) │
│ │
│ Reflect LAST WEEK's completed actions back as │
│ identity evidence. Not praise. Evidence. │
│ │
│ "You caught [X], solved [Y], handled [Z] before it │
│ escalated. That's your standard now." │
│ │
│ DO: Name specific actions. Connect to identity. │
│ DO NOT: Generic praise. Effort praise. Compliance. │
│ │
│ This is a MAINTENANCE OPERATION, not motivation. │
│ Without it, identity attachment decays. The self- │
│ sustaining cycle needs periodic evidence that the │
│ standard is real. Skip this and you will not notice │
│ the decay until week 4-5. │
│ │
└──────────────────────────────────────────────────────────┘
MONTHLY ADD-ON (10 min, first meeting of month):
┌──────────────────────────────────────────────────────────┐
│ │
│ EXTERNALIZATION CHECK │
│ │
│ "What standard did you write or teach this month?" │
│ │
│ Review the artifact. │
│ GATE: "Could a new hire use this without calling you?" │
│ YES → externalization is working. │
│ NO → knowledge is still tacit. They are the bottleneck. │
│ │
│ Grade portability, not quality. │
│ │
└──────────────────────────────────────────────────────────┘
MODE 5: AOM DEVELOPMENT
Separate from the weekly meeting. The meeting is a weekly cadence.
AOM development is the long arc. Where are they on the 13-week
trajectory? What is their current tier? What needs to happen
for them to advance?
AOM PERFORMANCE TIERS:
┌─────────────┬───────────────────────────────────────────────────┐
│ TIER │ SIGNALS │
├─────────────┼───────────────────────────────────────────────────┤
│ │ │
│ STRUGGLING │ - Primary metrics below target 2+ weeks │
│ │ - Asks permission instead of reporting decisions │
│ │ - Blank on Q1 (measurement routine broken) │
│ │ - Names symptoms, not constraints (Q3) │
│ │ - Escalation budget exceeded │
│ │ - Avoids difficult conversations with team │
│ │ - You catch problems they should have caught │
│ │ │
│ STABLE │ - Primary metrics near target, occasional misses │
│ │ - Reports most decisions, asks on some │
│ │ - Answers Q1-Q3 without prompting │
│ │ - Identifies constraints correctly ~60% of time │
│ │ - Self-corrects after flag (aware but slow) │
│ │ - Handles routine difficult conversations │
│ │ │
│ STRONG │ - Primary metrics consistently at/above target │
│ │ - Reports all decisions. Asks only on exceptions │
│ │ - Raises problems AND constraints AND solutions │
│ │ before you ask │
│ │ - Independently identifies and fixes constraints │
│ │ you did not flag │
│ │ - Handles difficult conversations proactively │
│ │ - Has externalized at least one standard │
│ │ - Spot audits confirm standards hold without you │
│ │ │
└─────────────┴───────────────────────────────────────────────────┘
GRADUATED EXPOSURE SCHEDULE:
┌──────┬──────────────────────────┬──────────────────────────────┐
│ WEEK │ STAKES LEVEL │ YOUR ROLE │
├──────┼──────────────────────────┼──────────────────────────────┤
│ │ │ │
│ 1-4 │ Shift-level decisions │ Available by text. │
│ │ Menu sub, staff swap, │ Debrief same day or at visit.│
│ │ small equipment issue. │ Present at location 2-3x/wk.│
│ │ │ │
│ 5-8 │ Day-level decisions │ Available by text. │
│ │ Vendor call, schedule │ Debrief at weekly meeting. │
│ │ change, team conflict. │ NOT present for the decision.│
│ │ │ Visit 1x/week. │
│ │ │ │
│ 9-12 │ Week-level decisions │ Check-in at meeting only. │
│ │ Food cost correction, │ NOT available between │
│ │ hiring rec, training │ meetings for this category. │
│ │ plan, standard change. │ Visit 1x/2 weeks. │
│ │ │ │
│ 13+ │ Month-level decisions │ Monthly review. │
│ │ Menu engineering, labor │ Exception handling only. │
│ │ model, vendor negotiate. │ Visit for calibration only. │
│ │ │ │
└──────┴──────────────────────────┴──────────────────────────────┘
SIGNALS:
- Asking permission at current level = not ready to advance. Hold.
- Reporting decisions at current level = ready for next level.
- Making mistakes at current level = expected. The mistake IS the rep.
Debrief it. Do not punish it.
INTERNALIZATION THRESHOLD (why week 13 is the target):
Four conditions must all be met for the cycle to self-sustain:
┌────────────────────────────┬──────────────────────────────────┐
│ CONDITION │ DIAGNOSTIC │
├────────────────────────────┼──────────────────────────────────┤
│ 1. Standard clear enough │ Can they state the standard in │
│ to self-enforce │ specific, time-blockable terms? │
│ │ "Run a good kitchen" = fail. │
│ │ "Walkthrough 10, 2, 6" = pass. │
├────────────────────────────┼──────────────────────────────────┤
│ 2. Enough reps to imprint │ 66+ days of feedback reps. │
│ │ 13 weeks = 91 days (buffer). │
├────────────────────────────┼──────────────────────────────────┤
│ 3. Right action easier │ Does the environment make │
│ than drift │ the standard the easy path? │
│ │ If following it is harder than │
│ │ drifting, they will drift. │
├────────────────────────────┼──────────────────────────────────┤
│ 4. Identity attached │ Do they say "that's not how I │
│ to the standard │ run my kitchen" when a standard │
│ │ breaks? Or do they shrug? │
│ │ Identity attachment = cognitive │
│ │ dissonance from violation. │
└────────────────────────────┴──────────────────────────────────┘
If any is missing, the system will not self-sustain:
- Standard unclear → rewrite until time-blockable
- Not enough reps → more time at current level
- Action cost too high → redesign the environment
- No identity attachment → more Phase 3 reflection
TRUST LAYER DIAGNOSTICS:
When coaching is not landing, diagnose which trust layer is broken:
┌──────────────────────────────┬────────────┬─────────────────────┐
│ SYMPTOM │ BROKEN │ REPAIR │
├──────────────────────────────┼────────────┼─────────────────────┤
│ Complies when you watch. │ CHARACTER │ Sustained group- │
│ Drifts when you leave. │ │ serving acts. Many │
│ │ │ over time. Not one │
│ │ │ speech. │
├──────────────────────────────┼────────────┼─────────────────────┤
│ Follows direction but cannot │ JUDGMENT │ Narrate reasoning │
│ explain why. Correct behavior│ │ on-site. Make WHY │
│ but no transfer to new │ │ visible. Struggling │
│ situations. │ │ tier visit protocol. │
├──────────────────────────────┼────────────┼─────────────────────┤
│ Respects you but ignores │ COMPETENCE │ Be right on big │
│ operational direction. │ │ calls. Track record │
│ Listens to character advice, │ │ not persuasion. │
│ ignores tactical guidance. │ │ │
├──────────────────────────────┼────────────┼─────────────────────┤
│ Meeting goes well, floor │ ALL THREE │ Start with on-site │
│ goes poorly. Classic gap │ (or none │ trust building. │
│ between meeting and reality. │ built yet) │ Mode 4 assumes │
│ │ │ trust exists. Fix │
│ │ │ trust first. │
└──────────────────────────────┴────────────┴─────────────────────┘
AOM MEASUREMENT CASCADE (teaching AOM to run their own daily review):
Week 13 goal: AOM runs the full negentropy cycle autonomously.
The cycle starts with NOTICE. If AOM is not measuring daily,
they cannot notice before you tell them. Mode 4 Q1 ("What did
you notice?") tests whether measurement is happening. But
nothing trains it unless you explicitly cascade Mode 2.
┌──────┬────────────────────────────────────────────────────────┐
│ WEEK │ AOM MEASUREMENT TRAINING │
├──────┼────────────────────────────────────────────────────────┤
│ │ │
│ 1-2 │ TEACH. On a visit, walk AOM through YOUR daily data │
│ │ review process. Show them which dashboards, which │
│ │ metrics, what the thresholds are. Do it together. │
│ │ "This is how I know the state of your kitchen before │
│ │ I ask you. I want you to know it before I ask." │
│ │ │
│ 3-4 │ AUDIT. AOM runs their own daily review. At the weekly │
│ │ meeting, ask: "Walk me through your data this week. │
│ │ What flagged?" Compare their review to yours. The gap │
│ │ is the training target. Narrate what they missed. │
│ │ │
│ 5-8 │ VERIFY. AOM runs independently. Mode 4 Q1 is the │
│ │ check. If they answer with data ("error rate spiked │
│ │ Wednesday, I traced it to the new closer"), the │
│ │ measurement routine is installed. If blank or vague, │
│ │ go back to week 3-4 mode. │
│ │ │
│ 9+ │ SELF-SUSTAINING. AOM's daily review feeds directly │
│ │ into their own predict-then-compare with shift leads. │
│ │ The cascade is complete when AOM is running Mode 4 │
│ │ with their shift leads using data from their own │
│ │ Mode 2. │
│ │ │
└──────┴────────────────────────────────────────────────────────┘
ESCALATION BUDGET (operationalized):
Definition: An ESCALATION is asking permission OR reporting
uncertainty at the AOM's current graduated-exposure level.
Reporting a completed decision is NOT an escalation.
Reporting an emergency is NOT an escalation.
┌──────┬──────────────────────┬──────────┬──────────────────────┐
│ WEEK │ DECISION LEVEL │ BUDGET │ WHAT COUNTS │
├──────┼──────────────────────┼──────────┼──────────────────────┤
│ 1-4 │ Shift-level │ 4/week │ "Should I sub the │
│ │ │ │ menu item?" = YES │
│ │ │ │ "I subbed it" = NO │
├──────┼──────────────────────┼──────────┼──────────────────────┤
│ 5-8 │ Day-level │ 2/week │ "Should I call the │
│ │ │ │ vendor?" = YES │
│ │ │ │ "I called them" = NO │
├──────┼──────────────────────┼──────────┼──────────────────────┤
│ 9-12 │ Week-level │ 1/week │ "Should I change │
│ │ │ │ the prep plan?" = YES │
│ │ │ │ "I changed it" = NO │
├──────┼──────────────────────┼──────────┼──────────────────────┤
│ 13+ │ Month-level │ 0.5/wk │ Only genuine │
│ │ │ (2/month)│ exceptions. │
└──────┴──────────────────────┴──────────┴──────────────────────┘
TRACKING (add to Friday close):
┌──────┬─────────┬──────────┬──────┬───────────┬────────────────┐
│ WEEK │ AOM │ BUDGET │ USED │ REMAINING │ TREND │
├──────┼─────────┼──────────┼──────┼───────────┼────────────────┤
│ │ Angel │ │ │ │ ↑ ↓ → │
│ │ Lizbeth │ │ │ │ ↑ ↓ → │
└──────┴─────────┴──────────┴──────┴───────────┴────────────────┘
TIER SIGNALS FROM ESCALATION DATA:
- Exceeded budget 2+ weeks running → hold current level or downgrade
- Below budget 4 weeks → consider tier upgrade
- Budget trending down → regulation building, exposure working
- Budget trending up → regression, investigate (stress? new challenge?)
MODE 5B: TRAINING AOM PROTOCOL (Angel, weeks 1-8)
Angel is in the 8-week graduation gate. This is not the same as
managing a stable or strong AOM. Training AOMs need a different
protocol because they are building the foundation, not refining it.
ANGEL'S CURRENT STATE (as of 2026-06-04):
Start date: 2026-05-18
Current week: ~2.5 of 8
Status: Training AOM (not graduated)
Location: Avondale
Comp: $55k → $65k on graduation
Reports to: Lydell Tyler (you)
THE TRAINING AOM IS ON-SITE BY DEFAULT.
Angel's offer letter requires daily inspections, floor monitoring,
prep process oversight, inventory counts, hiring, and
accountability logging. This is not an off-site manager role
during training. The off-site management pattern (lead through
data and systems) comes AFTER graduation when the systems are
built and proven.
During training, Angel should be on-site for most shifts where
he is scheduled as the manager on duty. The DM OS scheduling
model (2-3 on-site days) applies POST-GRADUATION when the
systems are running and the SL layer can hold the floor without
the AOM present.
GRADUATION GATE (8-week window, all must be met):
┌────────────────────────┬─────────────────────────────────────┐
│ AREA │ STANDARD │
├────────────────────────┼─────────────────────────────────────┤
│ Schedule drafted │ 100% published by Thursday night, │
│ │ two weeks out │
├────────────────────────┼─────────────────────────────────────┤
│ Schedule quality │ Built against forecast, role │
│ │ coverage, labor guardrails, │
│ │ known availability │
├────────────────────────┼─────────────────────────────────────┤
│ Task list on-time │ >= 90% completed on management days │
├────────────────────────┼─────────────────────────────────────┤
│ Admin action items │ >= 90% within 24hrs on mgmt days │
│ │ (<= 48hrs if non-scheduled day) │
├────────────────────────┼─────────────────────────────────────┤
│ Punch audit │ 100% reviewed within 24hrs of next │
│ │ scheduled mgmt day │
├────────────────────────┼─────────────────────────────────────┤
│ Star rating │ 30-day rolling >= lower of: 4.8 or │
│ │ location 90-day baseline. No single │
│ │ week -0.20 below without documented │
│ │ corrective action. │
├────────────────────────┼─────────────────────────────────────┤
│ Error rate │ 30-day rolling <= higher of: 0.75% │
│ │ or location 90-day baseline. No │
│ │ single week +0.50pp without │
│ │ documented corrective action. │
├────────────────────────┼─────────────────────────────────────┤
│ Prep time │ 30-day rolling <= higher of: 9 min │
│ │ or location 90-day baseline. No │
│ │ single week +1.5 min without │
│ │ documented corrective action. │
├────────────────────────┼─────────────────────────────────────┤
│ Food safety score │ >= 90% monthly. Critical issues │
│ │ corrected + documented within 24hrs │
├────────────────────────┼─────────────────────────────────────┤
│ Accountability logged │ 100% of attendance, conduct, food │
│ │ safety, service, performance events │
├────────────────────────┼─────────────────────────────────────┤
│ Escalations to DM │ <= 1 preventable non-urgent/week │
│ │ (emergencies, safety, HR, company- │
│ │ approval items don't count) │
├────────────────────────┼─────────────────────────────────────┤
│ Training completion │ 100%: all audits, ordering playbook,│
│ │ scheduling/punch/accountability/ │
│ │ food safety workflows signed off, │
│ │ PDP-zero conversation completed │
└────────────────────────┴─────────────────────────────────────┘
DM PROTOCOL DURING TRAINING PHASE:
WEEKS 1-2 (installing):
- Visit Avondale 2-3x/week. AOM-present visits.
- Narrate reasoning on every visit. This is Mode 3
struggling-tier behavior. Angel is learning HOW you
read the floor.
- Weekly meeting focuses on: "What did you notice this
week?" (Q1). If blank, the measurement walk is not
happening. Rebuild it before anything else.
- Check graduation gate metrics weekly. Not to grade.
To establish the baseline and show Angel what the
targets are in real data, not a document.
- Primary risk: Angel defaults to doing shift lead work
(working stations, filling gaps) instead of managing.
If you see this, name it: "You are filling a gap right
now. That is shift lead work. What would you tell
the shift lead to do instead?"
WEEKS 3-4 (testing):
- Drop to 1-2 visits/week. Mix AOM-present and
AOM-absent visits.
- Weekly meeting adds predict-then-compare (Mode 4
Phase 1). Start with 1 situation per meeting.
- Check: Is Angel identifying constraints (Q3) or
naming symptoms? If symptoms, narrate more on visits.
If constraints, increase graduated exposure.
- Check: Is the schedule being published on time?
Is the punch audit happening? These are the
administrative hygiene metrics that predict whether
the operational metrics will hold.
WEEKS 5-6 (graduating):
- 1 visit/week. Prefer AOM-absent.
- Weekly meeting runs the full 5-question protocol.
- P-T-C gap should be narrowing. If still above 40%,
Angel needs more on-site coaching reps.
- Start the externalization check: "Can you write down
how you do [X] so a shift lead could follow it?"
- Check escalation count. If above 1/week, the
regulation window is still narrow.
WEEKS 7-8 (verifying):
- 1 visit/2 weeks. AOM-absent only.
- The kitchen on an Angel off-day is the graduation test.
- All metrics must be at gate for 2 consecutive weeks.
- If any metric fails, the 8-week window extends.
- Prepare PDP-zero conversation: operating structure,
current constraints, improvement priorities, leadership
expectations, transition to full AOM.
POST-GRADUATION:
- Angel moves to standard AOM protocol (Mode 5).
- Monthly sustainability review against same gate metrics.
- Comp increases to $65k.
- On-site cadence shifts to the 2-day model (post-weekend
+ midweek peak). Off-site management begins.
- The transition from training to steady-state is the
highest-risk moment. Most regression happens in weeks
9-12 when the DM reduces oversight and the AOM has not
yet built the self-sustaining habit loop. Keep the
weekly meeting cadence tight through week 13 minimum.
MODE 6: CRISIS
An event that exceeds the AOM's current graduated-exposure level
AND has real consequences (financial, safety, team, customer).
Your behavior here programs their crisis response permanently.
Stress narrows attention. In that narrow window, their system
records your response at maximum encoding strength.
PROTOCOL:
1. REGULATE YOURSELF FIRST.
Pause. 30 seconds. Your nervous system sets theirs via
mirror circuit. If you call panicked, they panic.
2. ASK, DO NOT TELL.
"What do you know so far?"
"What have you tried?"
Even in crisis, their computation must run.
Override in crisis = they learn to freeze and wait for you.
3. IF THEY HAVE A PLAN, LET THEM RUN IT.
Even if yours would be better. A workable plan they
execute NOW > your optimal plan after a 20-min call.
4. IF THEY ARE FROZEN, GIVE ONE INSTRUCTION.
Not three. One. "Do [X]. Now. Call me after."
One preserves the gate. Three = full override.
5. DEBRIEF AFTER.
"What did you see? What would you do differently?"
This is the highest-quality predict-then-compare rep
available. Crisis reps encode at 20-50x normal because
arousal amplifies the error signal. Do not waste it on blame.
6. REFLECT AS IDENTITY EVIDENCE.
"You handled that. That's your standard now."
Crisis win + identity reflection = strongest identity
update possible. The next crisis activates this identity
instead of the freeze response.
DO NOT:
- Drive to the location unless safety is at risk.
Your presence says "I don't trust you to handle this."
- Text 15 questions. One call. One question at a time.
- Revisit the crisis repeatedly. One debrief. Move forward.
- Use the crisis as evidence of weakness. Use it as growth.
SELF-ASSESSMENT GATE
Run this every Friday during weekly close. 3 minutes.
This is your negentropy cycle for YOUR leadership.
WEEKLY GATES YES NO
1 Did I let them solve a problem I wanted □ □
to solve for them this week?
(Control Paradox)
2 Did I capture 3+ situations for □ □
predict-then-compare?
(Pattern library fuel)
3 Did I narrate my reasoning on a visit? □ □
(Mental model transfer / judgment trust)
4 Did I stay regulated when bad news arrived? □ □
(Mirror circuit quality)
5 Did I redirect a team member back to □ □
their AOM instead of solving?
(Authority reinforcement)
6 Did I reflect completed actions as □ □
identity evidence in the meeting?
(Winner effect / reinforcement operation)
7 Did I hold a data flag for the meeting □ □
instead of texting the AOM about it?
(Autonomy preservation)
8 Did I admit an error or uncertainty in □ □
front of them this week?
(Character trust)
9 Did every visit have a specific question? □ □
(Purpose-driven presence)
10 Did I do the daily data review every day? □ □
(DM measurement routine)
SCORING:
8-10 YES Operating as a leader of leaders. System building
toward self-sustaining. Your input is compressing.
5-7 YES Functional but leaking. Which NOs are habitual
(same ones every week)? Those are YOUR constraint.
0-4 YES Operating as a high-output individual. Your managers
are waiting for you, not computing for themselves.
Start with gate 1.
MONTHLY TRUST AUDIT (first Friday of month, add to weekly close):
11 Does Angel believe my operational calls □ □
are correct? (COMPETENCE)
12 Does Lizbeth believe my operational calls □ □
are correct? (COMPETENCE)
13 Do they believe I act in the group's □ □
interest? (CHARACTER)
14 Do they understand WHY behind my calls? □ □
(JUDGMENT)
Any NO on 11-14 = the weekly meeting will feel like
interrogation. Fix the trust layer on-site BEFORE
expecting the meeting to develop them.
QUESTION BANK
Daily Data Review
INTERNAL (to yourself, during data review):
"Is this a flag or noise? Would I flag this if it happened 3x?"
"Effort issue or system issue?"
"If I text the AOM about this, am I overriding or escalating?"
(Default: hold for meeting. Only escalate on severity override.)
WHEN SEVERITY OVERRIDE TRIGGERS (contacting AOM):
"What happened with [specific metric] yesterday?"
(One question. Let them explain. Then decide.)
Location Visit
PRE-VISIT (to yourself):
"What is the ONE question I am here to answer?"
"What tier is this AOM? How does that change my behavior?"
WALKING WITH AOM:
"Walk me through your floor." (their route, not yours)
"What are you seeing?" (tests signal detection)
"Where's your constraint today?" (tests model depth)
"What would happen if [X]?" (installs constraint thinking)
NARRATING REASONING (struggling AOM):
"I'm looking at [X] because when [X] is off, [Y] downstream."
"The reason I check this first is [mechanism]."
"Here's how I'd think about this: [reasoning chain]."
WHEN THEY ASK YOU TO SOLVE:
"What do you think?" (redirects computation)
"What's your read?" (tests pattern library)
"What would you tell your shift lead to do?" (leader framing)
WHEN YOU SEE SOMETHING THEY MISSED:
"I noticed [X]. What do you think that means?"
"Walk me through this station." (lets them discover it)
WHEN TEAM COMES TO YOU:
"Have you talked to Angel/Lizbeth about this?"
"That's Angel's/Lizbeth's call."
Weekly Meeting
PHASE 1 (predict-then-compare):
"This happened at [location] on [day]. What would you do?"
"You said [X] would fix [Y]. The numbers say [Z]. Why?"
"Before I tell you what happened, what do you think caused it?"
PHASE 2 (the 5 questions):
Q1: "Scan these: order errors, prep time, star ratings,
food safety, attendance, conduct. What did you notice?"
Q2: "What's the biggest problem?"
Q3: "What's the constraint?"
Q4: "What are you going to try?"
Q5: "When are you doing it?"
PHASE 2 (redirects):
"Upstream. What causes that?"
"If you could only fix one."
"What's the smallest version you could do today?"
"Not 'this week.' When specifically?"
"If we fixed that tomorrow, would the metric actually move?"
"Last week the constraint was [X]. Is it still [X]?"
PHASE 3 (identity reflection):
"You caught [X], solved [Y], handled [Z]. That's your standard."
"Your kitchen ran [metric] all week. That's not luck. That's you."
MONTHLY (externalization):
"What standard did you write or teach this month?"
"Could a new hire use this without calling you?"
Off-Shift (text/call)
WHEN THEY ASK PERMISSION:
"What do you think is the right move?"
"I trust your read. Handle it and tell me how it went."
WHEN THEY REPORT A DECISION:
"Good." (brevity IS the trust signal)
"How did it land?"
WHEN YOU WANT TO INTERVENE BUT SHOULDN'T:
Say nothing. Capture for predict-then-compare.
The urge to intervene is YOUR regulation test. (Gate 7.)
13-WEEK TRAJECTORY
WEEK WHAT IS HAPPENING WHAT YOU SEE
1-2 Installing measurement routines. AOM can describe "healthy"
Predict-then-compare gaps wide. for their kitchen in 3
5 questions require prompting. sentences. Gaps are 60%+.
Visits: 2-3x/week. Narrating. Daily data review installing.
3-4 Routines taking hold. Gaps shrinking. AOM answers Q1-Q3 without
Starting to identify constraints prompting. Gap shrinking
vs symptoms. MEASURE operation to 40%. Occasional correct
becoming habitual. constraint identification.
Visits: 2x/week. You catch fewer surprises.
5-6 Graduated exposure: day-level. AOM reports decisions
Identity language appearing. instead of asking. "I
Mirror circuit installing (you see handled it" becomes
your patterns in them). Visits: natural. Gaps at 30%.
1x/week. Asking, not narrating. First unprompted constraint ID.
7-8 Self-regulation expanding. AOM AOM handles a difficult
handles situations that would have conversation you expected
frozen them at week 1. Visits: them to avoid. First
1x/week. Questions only. externalized standard appears.
9-10 Transition to check-in mode. You have less to say in
Your input declining. Visits: meetings. Their answers
1x/2 weeks. Observe mostly. converge with yours.
Shared mental model forming. Gaps under 20%.
11-12 Spot audits confirm standards hold AOM raises problems AND
without your presence. Visits: solutions AND constraints
1x/2 weeks. Verification only. in Q1-Q3 before you ask.
Weekly meeting becomes verification.
13+ SELF-SUSTAINING. All four conditions AOM runs the full cycle:
met: clear standards, 66+ day reps, notice → analyze → solve →
right action easier than drift, execute → verify.
identity attached to standard. No external fuel needed.
Your role: calibration, exceptions, Visits: monthly or as needed.
and teaching them to run this process The cascade begins: they
with their shift leads. Your invest- run this with their team.
ment propagates one layer down at The system scales through
~85% fidelity without added cost. people who scale through people.
DM ROI TRACKING (evidence of value)
Every action in this OS produces evidence. A CEO, COO, or Director
of Operations should be able to open a dashboard and see: what this
DM did, what it produced, and the historical portfolio of impact.
Not "I had meetings." Proof.
THE DM VALUE CHAIN:
┌──────────────────────────────────────────────────────────────┐
│ │
│ DM INPUT (tracked) → BUSINESS OUTPUT (measured) │
│ │
│ 1 hr daily data review → Flags caught before AOM │
│ 1-3 visits/week → Situations captured │
│ 2 meetings/week → AOM gap trend (shrinking?) │
│ Graduated exposure → Escalation budget trend │
│ Identity reflections given → AOM tier transitions │
│ │
│ These inputs compound into: │
│ - Primary metric trends (error rate, ratings, prep time) │
│ - AOM development velocity (tier progression speed) │
│ - System self-sufficiency (DM intervention declining?) │
│ - Cross-location learning (transfer events per month) │
│ │
└──────────────────────────────────────────────────────────────┘
WHAT TO LOG (per mode):
MODE 2: DAILY DATA REVIEW
┌──────┬────────┬──────────────────┬──────────┬────────────────┐
│ DATE │ TIME │ FLAGS FOUND │ ACTION │ OUTCOME │
│ │ (min) │ (metric, loc) │ TAKEN │ (resolved?) │
├──────┼────────┼──────────────────┼──────────┼────────────────┤
│ │ │ │ HOLD / │ │
│ │ │ │ INVEST / │ │
│ │ │ │ ACT NOW │ │
└──────┴────────┴──────────────────┴──────────┴────────────────┘
WEEKLY METRIC: Data review completion rate (target: 5/5 weekdays)
MODE 3: LOCATION VISIT
┌──────┬─────────┬──────────────────┬──────────────────────────┐
│ DATE │ LOC │ QUESTION │ ANSWER FOUND │
│ │ │ (pre-visit) │ │
├──────┼─────────┼──────────────────┼──────────────────────────┤
│ │ │ SITUATIONS │ AOM BEHAVIOR │
│ │ │ CAPTURED (#) │ (narrated/questioned/ │
│ │ │ │ observed) │
└──────┴─────────┴──────────────────┴──────────────────────────┘
WEEKLY METRIC: Visits completed vs plan. Situations captured.
MODE 4: WEEKLY MEETING
┌──────┬─────────┬──────────────┬──────────┬───────────────────┐
│ DATE │ AOM │ P-T-C GAP │ Q3 HIT? │ ID REFLECTION │
│ │ │ (% this wk) │ (binding │ GIVEN? (Y/N) │
│ │ │ │ found?) │ (what evidence?) │
├──────┼─────────┼──────────────┼──────────┼───────────────────┤
│ │ │ │ │ │
└──────┴─────────┴──────────────┴──────────┴───────────────────┘
WEEKLY METRIC: P-T-C gap trend. Constraint identification accuracy.
MODE 5: AOM DEVELOPMENT
┌──────┬─────────┬──────────┬──────────┬──────────┬───────────┐
│ DATE │ AOM │ TIER │ ESC. │ ESC. │ NOTES │
│ │ │ (S/St/Sg)│ BUDGET │ USED │ │
├──────┼─────────┼──────────┼──────────┼──────────┼───────────┤
│ │ │ │ │ │ │
└──────┴─────────┴──────────┴──────────┴──────────┴───────────┘
MONTHLY METRIC: Tier transitions. Escalation trend direction.
MODE 6: CRISIS
┌──────┬─────────┬──────────────┬──────────┬───────────────────┐
│ DATE │ LOC │ EVENT │ AOM │ DEBRIEF │
│ │ │ │ RESPONSE │ COMPLETED? (Y/N) │
│ │ │ │ (froze/ │ GAP IDENTIFIED? │
│ │ │ │ plan/ │ │
│ │ │ │ solved) │ │
└──────┴─────────┴──────────────┴──────────┴───────────────────┘
SELF-ASSESSMENT GATE
┌──────┬────────────────────────────────────────────────────────┐
│ WEEK │ SCORE (out of 10) │
├──────┼────────────────────────────────────────────────────────┤
│ │ Gate 1: □ Gate 2: □ Gate 3: □ Gate 4: □ Gate 5: □│
│ │ Gate 6: □ Gate 7: □ Gate 8: □ Gate 9: □ Gate 10:□│
│ │ TOTAL: __/10 │
└──────┴────────────────────────────────────────────────────────┘
MONTHLY METRIC: Self-assessment trend. Which gates fail habitually?
DM PORTFOLIO DASHBOARD (what the CEO sees):
┌──────────────────────────────────────────────────────────────┐
│ DM SCORECARD │
│ │
│ LEADING INDICATORS (DM activity) │
│ ┌─────────────────────────┬─────────┬──────────┬──────────┐ │
│ │ METRIC │ TARGET │ THIS WK │ TREND │ │
│ ├─────────────────────────┼─────────┼──────────┼──────────┤ │
│ │ Data review completion │ 5/5 │ │ ↑ ↓ → │ │
│ │ Visits completed │ per plan│ │ ↑ ↓ → │ │
│ │ Meetings held │ 2/2 │ │ ↑ ↓ → │ │
│ │ Situations captured │ 3+ │ │ ↑ ↓ → │ │
│ │ Self-assessment score │ 8+/10 │ │ ↑ ↓ → │ │
│ └─────────────────────────┴─────────┴──────────┴──────────┘ │
│ │
│ LAGGING INDICATORS (business results) │
│ ┌─────────────────────────┬─────────┬──────────┬──────────┐ │
│ │ METRIC │ TARGET │ THIS WK │ TREND │ │
│ ├─────────────────────────┼─────────┼──────────┼──────────┤ │
│ │ Error rate (S. Loop) │ <0.9% │ │ ↑ ↓ → │ │
│ │ Error rate (Avondale) │ <0.9% │ │ ↑ ↓ → │ │
│ │ Star rating (S. Loop) │ 4.8+ │ │ ↑ ↓ → │ │
│ │ Star rating (Avondale) │ 4.8+ │ │ ↑ ↓ → │ │
│ │ Prep time (S. Loop) │ target │ │ ↑ ↓ → │ │
│ │ Prep time (Avondale) │ target │ │ ↑ ↓ → │ │
│ │ Labor % (both) │ 17.5% │ │ ↑ ↓ → │ │
│ └─────────────────────────┴─────────┴──────────┴──────────┘ │
│ │
│ DEVELOPMENT INDICATORS (team building) │
│ ┌─────────────────────────┬─────────┬──────────┬──────────┐ │
│ │ METRIC │ TARGET │ CURRENT │ TREND │ │
│ ├─────────────────────────┼─────────┼──────────┼──────────┤ │
│ │ Angel tier │ Strong │ │ ↑ ↓ → │ │
│ │ Lizbeth tier │ Strong │ │ ↑ ↓ → │ │
│ │ P-T-C gap (Angel) │ <10% │ │ ↑ ↓ → │ │
│ │ P-T-C gap (Lizbeth) │ <10% │ │ ↑ ↓ → │ │
│ │ Escalation trend │ ↓ │ │ ↑ ↓ → │ │
│ │ Cross-loc transfers │ 1+/mo │ │ ↑ ↓ → │ │
│ │ Standards externalized │ 1+/mo │ │ ↑ ↓ → │ │
│ │ DM intervention rate │ ↓ │ │ ↑ ↓ → │ │
│ └─────────────────────────┴─────────┴──────────┴──────────┘ │
│ │
│ TRUE ROI: │
│ DM intervention rate declining + primary metrics improving │
│ = system is becoming self-sustaining = DM can scale to │
│ more locations without proportional time increase. │
│ │
│ The ultimate proof of DM value: │
│ Week 1: DM works 6 hrs/week on 2 locations. │
│ Week 13: DM works 4 hrs/week on 2 locations, metrics same. │
│ Week 26: DM adds location 3, total hours = 5. │
│ The time per location DECREASES as AOMs self-sustain. │
│ That is the ROI. Not "I was busy." "I became unnecessary." │
│ │
└──────────────────────────────────────────────────────────────┘
DATABASE SCHEMA (for implementation):
dm_activity_log:
id, date, mode (2/3/4/5/6), location, duration_min,
description, flags_found, action_taken, outcome
dm_meeting_log:
id, date, aom, ptc_gap_pct, constraint_found (bool),
constraint_binding (bool), identity_reflection (text),
q4_action, q5_when
dm_visit_log:
id, date, location, aom_tier, pre_question, answer,
situations_captured (int), aom_behavior_notes
dm_escalation_log:
id, date, aom, decision_level, was_escalation (bool),
description, outcome
dm_self_assessment:
id, week_start, gate_scores (json: {g1: bool, ... g10: bool}),
total_score, habitual_failures (text)
dm_aom_tier_history:
id, date, aom, tier (struggling/stable/strong),
evidence (text), transition_direction (up/down/hold)
dm_weekly_metrics:
id, week_start, location,
error_rate, star_rating, prep_time, labor_pct,
dm_data_review_count, dm_visit_count, dm_meeting_count,
dm_situations_captured, dm_self_assessment_score
THE SINGLE PRINCIPLE
The goal is not to make them listen to you.
The goal is to make your thinking unnecessary by installing a copy of it in their system.
Shared mental model + habituated measurement + identity attachment = autonomy.
4-6 hours per week to run two ghost kitchens through two managers. When the output runs without your input, you have built leaders, not followers. And then the real leverage appears: they cascade the system one layer down.