Inspectable TID / Market Gaps material. No interpretation is generated from this file. public_test

Kitchen-pilot protocol

Protocol for a bounded kitchen test.

Collection Operations and pilot design Type MARKDOWN Size 8453 bytes Original /assets/market-gaps/100-burger-a-day/04_PILOT/kitchen_pilot_001_protocol.md

Download original

Markdown preview

391 lines

Gate 11E — Kitchen Pilot 001

Status: READY TO RUN

Purpose: replace modeled seconds with observed seconds using 20 real premium combos.

This is not a recipe-development day and not a sales event. It is an engineering pilot.

Pilot output

Sale tokens

20 test sale tokens
each token represents one customer order in the pilot

Non-sale units

1 QC patty/burger allowance
1 remake/recovery patty allowance

Physical patties prepared:

22 × 140 g = 3.08 kg raw premium blend

Only 20 are test-sale units. The two extras are never treated as orders #21–22.

Reference combo

For Kitchen Pilot 001, deliberately keep the menu minimal:

one premium beef burger
one bun
one standard sauce
one pickle/green configuration
standard cheese only if cheese is already part of the chosen reference burger
150 g finished tallow fries
one house soda

No build-your-own customization.

The purpose is to validate the operating system, not menu breadth.

Required ingredients

Beef

22 × 140 g premium patties
total raw blend: 3.08 kg

Buns

22 buns minimum
20 test combos
1 QC
1 spare/remake

Fries

Target sold output:

20 × 150 g = 3.0 kg finished fries

Buy/prep enough potato product to comfortably exceed this after process loss. Record:

raw/pre-cook weight
finished served weight
waste
leftover usable product

Do not infer potato yield from the target. Measure it.

Soda

minimum 22 serving equivalents
record actual ml/serving
record syrup and water usage if postmix/house syrup is used

Condiments / packaging

Prepare enough for 22 units but count actual use.

Minimal equipment

Required:

flat-top / plancha or equivalent test surface
fryer suitable for the selected fat
thermometer(s)
digital scale
timer/stopwatch
cold storage
raw-meat tools
cooked-only tools
fry baskets/tools
assembly surface
ticket/order display method
20 numbered order cards
2 reserve/QC cards
waste container
data sheets / phone/tablet for timestamps

Useful but not mandatory:

thermal probe logger
second camera for later motion review
separate soda dispenser
heat lamp/pass shelf if testing short hold

Layout rule

Before cooking:

mark RAW zone
mark HOT/COOKED transfer zone
mark FINISH zone
mark HANDOFF zone
place all frequently used items within one-step reach

No raw-contact utensil may enter the finish zone.

Pilot roles

Operator A — HOT

patties
cheese if standard hot-side step
fries
cooked transfer
hot-side temperature checks

Operator B — FINISH

bun
sauce/toppings
patty receive
fries portion
soda
ticket reconciliation
handoff

Observer / recorder

Strongly preferred. The observer does not rescue the workflow unless a safety problem occurs.

If only two people are available:

record with fixed video
log only essential timestamps live
annotate detail from video afterwards

Order release schedule

The pilot has three phases.

Phase 1 — Warm-up

5 orders at 20/hour

one new order every 3 minutes

Release:

#01 at 00:00
#02 at 03:00
#03 at 06:00
#04 at 09:00
#05 at 12:00

Purpose:

verify station flow
catch obvious layout problems
avoid learning the system for the first time during rush

Phase 2 — Forced rush

10 orders at 50/hour

one new order every 72 seconds

Release:

#06 at 15:00
#07 at 16:12
#08 at 17:24
#09 at 18:36
#10 at 19:48
#11 at 21:00
#12 at 22:12
#13 at 23:24
#14 at 24:36
#15 at 25:48

Purpose:

stress the exact reference peak
test whether HOT and FINISH can keep flow without long holding

Phase 3 — Recovery

5 orders at 25/hour

one every 2 minutes 24 seconds

Release:

#16 at 28:00
#17 at 30:24
#18 at 32:48
#19 at 35:12
#20 at 37:36

Purpose:

see whether the system recovers after rush
detect accumulated grease, fryer recovery problems, stock-positioning problems or operator fatigue

Ticket flow

Reference simulation:

customer order token is released
ticket becomes visible to kitchen
sale-token count decrements
kitchen begins work
ready timestamp recorded
ticket number called
handoff timestamp recorded

Public counter for the test: 20 → 0

Reserve/QC units do not affect the public counter.

Patty batch test

Start with the current reference hypothesis:

5 patties per rolling batch
new batch approximately every 6 minutes during the rush

Do not force this cadence if the physical cook curve proves it wrong.

For every batch record:

load time
flip time
cheese time if used
transfer time
surface temperature before/after load if measurable
first and last patty ready times
any hold duration

Fries batch test

Start with:

~1.25 kg finished-output target per batch
approximately one batch every 10 minutes at 50-combo/hour equivalent

For every batch record:

pre-cook/raw input weight
load time
oil/tallow temperature before load
temperature recovery
lift time
drain time
finished weight
waste
crispness score at 0 / 5 / 10 min if possible

QC tasting

Before order #01:

taste sauce
taste first fries
taste at least one representative cooked patty sample
verify salt/seasoning
verify bun condition

If a full QC burger is made, code it: QC-01

It is not test sale #01.

Mandatory measurements per order

For each of the 20 sale tokens record:

order release time
hot start
patty ready
finish start
fries ready
soda ready
combo ready
handoff
any remake/error
any hold time
finish active seconds if observable
customer-ready latency

Sensory scorecard

For selected burgers, score 1–5:

crust
juiciness
seasoning
patty temperature
bun integrity
sauce balance
fries crispness
fries salt
overall burger
overall combo

At minimum taste:

one warm-up burger
one early-rush burger
one late-rush burger
one recovery burger

This is specifically to detect quality drift during peak.

Waste ledger

Every discarded or diverted item gets one code:

QC
REMAKE
DROP
OVERCOOK
UNDERCOOK
ASSEMBLY_ERROR
FRIES_WASTE
SODA_WASTE
OTHER

Record both count and weight where practical.

PASS / FAIL thresholds

Safety — HARD FAIL

Immediate FAIL if:

raw/cooked cross-contact occurs and is not contained;
validated cook endpoint is missed;
cold/hot handling falls outside the chosen documented process;
operator cannot maintain safe workflow under rush.

Throughput — PASS

For the 10-order rush:

all 10 orders must be absorbed without uncontrolled queue growth;
projected station capacity must be >=50 combos/hour;
90th-percentile customer-ready latency target: <=8 min

Finish/front — PASS

median active finish time: <=60 s/combo
90th percentile: <=75 s/combo

HOT — PASS

no systematic crust degradation during rush
no repeated surface-recovery failure
no need for long pre-cooking just to survive the queue

Fries — PASS

peak production sustains required output
no repeated starvation of assembly
quality remains acceptable through rush

Quality — PASS

No more than one full point average decline (1–5 scale) between:

and

warm-up quality sample
late-rush quality sample

Any repeated 2+ point deterioration triggers redesign.

Waste — provisional target

zero preventable remakes preferred
<=1 recovery/remake unit in 20 orders
total edible waste recorded, not hidden

Human factors — PASS

Each operator scores after pilot:

fatigue 1–5
stress 1–5
station clarity 1–5

A technically fast workflow that requires sustained 5/5 stress is not a PASS.

Output after pilot

Kitchen Pilot 001 produces:

observed median/p90 order latency
observed finish active time
observed HOT active time
actual patty batch cycle
actual fries yield and batch cycle
actual reserve usage
actual edible waste
actual operator stress/fatigue
sensory drift across service
revised Day 001 model

Decision rule

Do not optimize everything at once.

After the pilot, identify the single largest constraint:

HOT
fries
finish
ordering
layout
ingredient prep
cleaning
quality hold
human fatigue

Then change one major variable before Kitchen Pilot 002.

That keeps the experiment interpretable.