L1
Customer
What we believe the customer cares about most, wants most, needs most at this stop. A belief or need — never an action.
3 defined18 partial3 deliberately none
+3 defined this session
Twenty-three stops, four gates and one exit, in canonical order — scored against the six layers a stop needs before it is real. Click any cell for its definition, deliverables, acceptance and failure mode.
LC is now its own column. The conversion lever — the specific thing that delivers what the customer wants and moves them to the next stop — was previously buried and therefore skipped. As a column, an undefined lever is visible across the whole chain at once.
The discipline that keeps L1 and LC distinct: L1 is a belief or need and is never an action. LC is always an observable event. If you cannot watch LC happen, it is not an LC — which is also what makes L4’s completion rate computable.
B — Built — added 2026-08-19. The five thinking layers can all read green while a store still has nothing to use. B scores whether an operating surface actually exists for the seat that performs the stop. It is the only column that answers “is this in production,” and it is deliberately the last one — a surface built before its row is defined is the most expensive artefact in the programme.
What we believe the customer cares about most, wants most, needs most at this stop. A belief or need — never an action.
The specific thing that provides it and progresses them to the next stop. Always an observable event.
The deterministic and agent structures that present options autonomously, interactively, immediately — leveraging data science, cognitive load, choice architecture.
The human involvement required to accomplish the experience and the conversion objective.
Every measure for this stop: customer, intelligence effectiveness, human effectiveness, and LC completion rate.
Whether an operating surface actually exists for the seat that performs this stop, and how far it sits from a store using it. Every stop needs one — there is no such thing as a surfaceless stop.
Each cell opens five fields: definition, deliverables, acceptance, governs, and breaks if absent. L4 is scored against what the reporting eco-system map actually documents, including its own [TBD] markers — a stop counts as measured only where a named dashboard covers it. B is scored against the active project tree — defined means deployed to a store and in use, which nothing yet is.
| Behavioural spineStop | L1Customer | LCConversion | L2Intelligence | L3Human | L4Measurement | BBuilt |
|---|---|---|---|---|---|---|
| Segment 1Pre-Interactions | ||||||
| 01Web-Visits | ||||||
| 02Log-ins | ||||||
| 03Profiles | ||||||
| ◆ Gate A1 — Pre → Digital | ||||||
| Segment 2Digital | ||||||
| 04Leads | ||||||
| 05Connected | ||||||
| 06Appt-Set | ||||||
| ◆ Gate A2 — Digital → Physical | ||||||
| Segment 3Physical | ||||||
| 07Check-in | ||||||
| 08Discovery | ||||||
| 09MCP | ||||||
| 10DemoU1 | ||||||
| 11Write-Up | ||||||
| 12Deal | ||||||
| E1Left-No-Deal | ||||||
| ◆ Gate A3 — deal_created | ||||||
| Segment 4Deal Lifecycle — the only complete L4 block | ||||||
| 132-Min | ||||||
| 14Submitted | ||||||
| 15Approved | ||||||
| 16PaperedU2a | ||||||
| 17WalkedU3 | ||||||
| 18Bank-Pack | ||||||
| 19FundedU2b | ||||||
| 20Finalized | ||||||
| ◆ Gate A4 — Finalized → Ownership | ||||||
| Segment 5Ownership — new, largely undefined | ||||||
| 21First-Tripnorth star | ||||||
| 22Serviced | ||||||
| 23Referral | ||||||
Every cell is clickable. Gates carry LC and L4 only — a gate is a conversion, not a state. If the side panel does not open — on a phone preview, or anywhere scripts are blocked — every cell is written out in full under Cell detail below.
All of Segment 4 has a defined LC — submitted, approved, signed, walked, funded, closed. Those stops are events, so the lever names itself.
Every gap sits in Segments 1–3 and 5, where the conversion is a behaviour rather than a transaction. Six are missing outright: A1 · 04 · A2 · A4 · 21 · 23.
Two of those six are the most valuable conversions in the business. A2 — showing up: a booked appointment becoming a physical arrival, the largest leak in Segment 2, and the journey map itself flags show rate as not surfaced. And 21 — first trip completed, which is the North Star.
So the chain can describe everything about closing a deal, and almost nothing about getting someone through the door or getting them camping.
No stop scores defined on Built. Eleven have a prototype — the portal app, the Floor Workspace check-in wizard, six competing MCP surfaces, a desking prototype, a walkthrough capture tool, and the internet-lead experiment harness. Thirteen have no surface at all.
The inversion is the finding. Segment 4 is the only fully measured block in the chain and it contains zero surfaces — 13 through 20 are entirely unbuilt. Segment 3, the least finished on measurement, holds almost every prototype that exists. So the chain measures most precisely exactly where nobody can do anything, and is built most heavily where the numbers are thinnest.
Stop 17 remains the sharpest single cell: every metric exists, the CIT Tracker covers it, it carries a universal requirement — and it has no screen. Stop 21, the North Star, has no surface and no capture event.
Three L1 answers assigned four of the eight agents without anyone doing an L2 pass. Signal Detection and Clarity fell out of 01; Information Collection and Decision Support fell out of 02.
This is why L1 goes first even though L2 is the larger gap. L2 cannot be specified before L1 says what needs understanding — an agent with nothing to present is just a name.
The roster is fully specified — goal, bounded authority, tools, feedback loops. Twenty-one stops have no agent mapped, so every “the system handles this” in the fulfillment column remains an intention.
Every stop has exit criteria. Eighteen still do not define what the customer needs to understand. The hesitation ledger is the only pre-existing content and it clusters on 08–11 and E1.
Segment 4 is fully measured and barely built. Stop 17 is the sharpest case: the CIT Tracker is its primary dashboard, every metric exists — and it has no screen, no chunk, while being a universal requirement.
Segment 2 is the reverse — with two corrections as of 2026-08-19. The dashboards reaching 04 and 05 count leads rather than qualifying, pricing or timing them \u2014 which is why 04’s L4 was rescored partial on 2026-08-19. And these stops are not surfaceless: the internet-lead-to-appointment experiment harness runs both, emitting server-side events already. It is a harness, not a product — but “no surface at all” was wrong.
Gates A1 and A4 have no metric and no dashboard. E1 has zero measurement. Segment 5 has one partial L4 across three stops. Every one is a place a customer enters, leaves, or stops being counted.
| Layer | What it takes | Remaining |
|---|---|---|
| L1 | Define what the customer needs to understand at each stop, and in what form. Three answered; the intake file carries the open question for each of the rest. | 18 |
| LC | Name the observable event that moves the customer out of each stop. Must be watchable — if it cannot be observed it cannot be a lever, and L4 cannot rate it. Priority: A2 show rate and 21 first-trip completion. | 13 |
| L2 | Assign the eight agents to stops, then define what the system decides, presents, and hands to a human. Four agents are already in play from the L1 work. | 21 |
| L3 | Seat stops 21–23, answer “who owns Finalized,” then add the justification field to the 17 that already carry a role and verb. | 3 seats |
| L4 | Instrument A1, A2, A4, E1, 21, 23, and consume vw_survey at 22. Stop 21 needs a first-trip completion event that exists nowhere. | 6 + 1 event |
| B | Ship something. For the eleven prototypes: pick a candidate of record, close prototype-to-store-ready, benchmark first per R12, then roll out with fallback and rollback criteria. For the thirteen with nothing: name the seat and the device before drawing anything. Priority is where B and L4 disagree most — 17 and Segment 4, fully measured and entirely unbuilt. | 24 |
Every cell’s full definition, in reading order — the same content the side panel shows, rendered as plain HTML. This exists so the file still works where scripts do not run: a phone’s file preview, Quick Look, an email client, a print pipeline. Tap a stop to open it. Generated from the same data set as the matrix, so the two cannot disagree.
As the system, with no human present, I render one low-effort decision for an anonymous visitor and emit guide-started and guide-completed so a completion rate exists per traffic source.
Only stop where the customer is unknown and unobligated — a different state from 02, not a smaller one. Answered: "that they can figure this out." They must believe the process is manageable, safe and within their control before inventory or differentiators mean anything.
A cold visitor completes one micro-decision within 30 seconds, with no identity request appearing before it.
P1P2P3P4F1F2
Page defaults to inventory-first, landing "RVs fit your life" before anxiety drops. Premature commitment, abandonment. Attribution: system — no human is present to blame.
Completion of one micro-decision — the two-minute guide. Not dwell time, not pages viewed. A single self-guided act, because P3 says competence is proven by doing, never by being told.
02 leverGuide-completion rate is reportable per traffic source.
P3F1
No lever exists, so the stop is measured by proxies (time on site, bounce) that cannot be acted on. Attribution: process.
Two agents assigned as a byproduct of the L1 answer. Signal Detection supplies "what we know" for an anonymous visitor — entry source, pages viewed, returning-visitor state, geography, no identity. Clarity owns the framework and its one constraining question.
The greeting renders differently for a first-time and a returning anonymous visitor, with no identity involved.
F2P4
Greeting is generic, so F2 fails and the opening reads as a template. Attribution: system.
Deliberately none. Normalization is delivered by content, not a person. A human here creates the obligation P1 exists to avoid.
No human touch occurs before the customer chooses one.
P1P4
Adding a human here — chat pop-up, callback offer — reintroduces obligation and suppresses guide completion.
Journey map lists metrics but [TBD-P3] for dashboards — Web Listings, External Factors, Rates, Weather, Campgrounds in the Marketing cluster. The L1 answer changed the success signal to micro-decision completion, which nothing measures.
Guide-completion rate is visible per store and per source, and no one is reporting dwell time as success.
Type A
Segment 1 stays invisible to leadership — the largest population, unmeasured.
Prototype only — not deployed. Surfaces that exist: Customer Buying Portal and Customer Buying Portal App prototypes (roughly 100 HTML/JSX files between them), plus the “shopping online — first customer buying environment” concept boards. The live public site is not this surface. Nothing built renders the two-minute guide that the L1 answer and the LC lever both depend on.
Deployed to at least one pilot store and in daily use by the seat that owns the stop.
R12
A prototype gets mistaken for a capability. The roadmap assumes it is done, the benchmark is never taken, and the first real customer meets an unfinished surface. Attribution: process.
As the system, I offer to save the work the visitor has just created, requesting a phone number only after value has been delivered, and I never gate before the guide is complete.
The capability boundary, not a content stop. Answered: they become the author of their own Buying Playbook. What they gain is authorship, not content — and everything authored persists, follows them across 27+ stores, and is shared with whoever helps them.
A customer can say what they got by logging in, and it is not "an account."
P4P5
Login reads as a lead-capture gate. Guide-to-save conversion collapses. Attribution: system.
Saving the playbook — phone entered, magic link followed. Framed as preserving an asset they already built (P5), never as registering.
Guide-to-save conversion is reportable, and the prompt never uses the word "register."
P4P5
Customers who completed the guide lose it and leave. The most expensive abandonment in the funnel, because it happens after we earned trust.
Three more agents assigned. Information Collection owns the vault and every object request. Decision Support owns towing pass/fail, the 27-store pre-filter, and the payment slider. Clarity owns the Camping Vision framing. Four of eight now mapped.
customer_nodeA pasted competitor URL returns a normalized unit scored against the customer’s saved payload margin.
R22R38
Login saves a static list a cookie could replicate. The upgrade is storage, not capability — and the P5 promise is hollow.
Deliberately none at the save point. But this stop creates a hard L3 requirement downstream: the outfitter reads the authored profile rather than re-asking it.
07 and 08 acceptanceNo outfitter re-asks a question the customer already answered in their playbook.
P5
Re-asking destroys the authorship premise. The customer learns their work does not travel. Attribution: person.
[TBD-P1/P3] — Usage Dashboard covers post-purchase tool usage; the log-in funnel itself is uncovered. Metrics on file are logged-in sessions and return-visit rate, both vanity.
Two rates exist and are separately visible: guide-to-save, and authoring actions per saved customer.
Type A
We cannot tell whether the value-first pivot works, so the wall placement is unfalsifiable.
Prototype only — not deployed. Surfaces that exist: Customer Buying Portal App carries profile-creation and entry screens. The staged identity ladder — device ID → phone OTP / magic link → email → name, each rung paid for with value first — is specified in the whiteboard corpus and built nowhere.
Deployed to at least one pilot store and in daily use by the seat that owns the stop.
R12
A prototype gets mistaken for a capability. The roadmap assumes it is done, the benchmark is never taken, and the first real customer meets an unfinished surface. Attribution: process.
As the system, I collect trade, payoff, registration ZIP, credit tier and cash-vs-finance posture — only because payment math is impossible without them — and I return a passwordless path back in.
The context anchor — distinct from 02’s identity anchor. Answered: continuity, preservation, passwordless access — stated explicitly, never implied. An implied promise creates a clarity deficit that triggers resistance.
The promise is visible on screen at the moment of the ask, and no gate exists that cannot explain itself computationally.
P5P6
Generic "Register" button. Reads as a lead wall built for the dealership. Attribution: system.
Camping Vision Worksheet plus tow specs submitted. Distinct from 02’s save: this is context, not identity, and it unlocks the four capabilities the math cannot run without.
Profile-completion rate is reportable, and a customer can complete it in stages without losing progress.
P6
Login rate and profile rate collapse into one number, hiding which wall is actually leaking.
A required behaviour is specified though no agent owns it. Graceful prompt state: every profile-gated capability explains why computationally — "we need your truck to calculate this safely," never "complete your profile." R38 validation state applies: numbers declare estimated · pending · verified.
Every gated capability names the missing input and why it is required. Every provisional number shows its state.
P6R38R3
Gates read as extractive. Provisional numbers pass as settled, and the number changes at the finance desk.
Deliberately none — the profile is customer-authored. Passwordless SMS makes the phone both identity and auth factor.
A customer who changes their phone number does not lose their playbook.
Reassigned numbers become an account-takeover surface, and SMS failure silently blocks all access.
[TBD-P3] — likely surfaces in Marketing. Metrics on file: profile completions, profile-to-lead conversion. Two distinct rates are required, not one.
01→02) and profile rate (02→03) reported separatelyA leader can say which of the two walls is leaking without inference.
Type B
A leak at the identity wall and a leak at the context wall look identical, so the wrong one gets fixed.
Prototype only — not deployed. Surfaces that exist: Customer Buying Portal App profile, towing and trade screens. The passwordless return path the L1 promise states outright — a secure link, any device, no password — has no built implementation.
Deployed to at least one pilot store and in daily use by the seat that owns the stop.
R12
A prototype gets mistaken for a capability. The roadmap assumes it is done, the benchmark is never taken, and the first real customer meets an unfinished surface. Attribution: process.
Profile-to-lead conversion — the segment crossing from anonymous behaviour to a known, worked lead. No lever is defined.
The crossing is a named, observable event.
Segment 1 and Segment 2 are separately measured and unconnected. Nobody owns the handoff.
No throughput metric and no dashboard. Journey map: [TBD-P3], likely Marketing; no Sales dashboard reaches it.
The gate has a rate on the board.
Type A
The first segment boundary in the chain is invisible.
As the BDC rep or Outfitter, I ask for and capture the minimum identity on a new lead so that it dedupes to one customer, enters the cadence, and starts a clock I can be held to.
Boundary only. Exit criteria exist; what the customer needs to understand here does not.
What must the customer understand about what happens next — and how fast do we commit to it in writing?
A cold customer can state, in their own words, the one thing this stop was meant to make clear.
P1P3F1
Surface defaults to inventory-first or form-first. Customer proceeds without comprehension, then stalls later at a stop that assumed it. Attribution: system.
The lever is almost certainly a committed next step, in writing, fast — but the commitment has not been set, so there is nothing to instrument.
A stated time appears on screen and the same number is enforced internally.
R13R18
A lead sits with no promise made and no clock running. Attribution: process.
No agent assigned. The eight-agent roster is fully specified — goal, bounded authority, tools, feedback loops — and none is mapped here. "The system handles this" is an intention.
A customer can complete this stop with no human present, or the system states plainly why a human is required.
P6
The stop runs on human memory. Same customer gets a different experience per rep, per store, per day. Attribution: process.
First stop where a human enters the journey. Role: BDC or Outfitter, configurable per store. Verb: ask. Exits when the lead exists and is deduped to a customer.
customer_node on phone or emailEvery lead has a named owner within the stated commitment window.
R33
Leads are worked by whoever notices. Attribution stamps are missing, so nobody is measurable. Attribution: person.
Downgraded from defined 2026-08-19. Deals Created (Sales Ops) is the earliest top-of-funnel signal and Outfitter Statistics carries rep credit — cy_deals per day, PY/PY2, Growth %, Voided/Inactive %. That is lead volume and rep attribution. It is not what this stop asks. This stop asks three questions and none of them is answered: how effective are we at converting to a real lead (rather than someone trading a form for a price), what does a lead cost, and how fast do we get from lead to connected.
Lead Source. Cost is specified in the whiteboard corpus (invested-$ per phase, investment-per-lead $/#) and has no data feed — the one genuinely blocked item05, where the dashboard does not exist)A store can see, for its own leads: what share became a real two-way interaction, what each one cost, and the distribution of time-to-first-response — not just how many arrived.
Type AR12
The stop reports arrivals and calls it performance. Spend is unattributable, so nobody can tell a channel that delivers buyers from one that delivers form-fills, and the response-time promise at L1 cannot be made because the current distribution is unknown. Attribution: process.
Prototype only — not deployed. Surfaces that exist: internet-lead-to-appointment.build — a working experiment harness with a routing server, cadence engine, append-only event emitter and an operator console — plus 08-Internet-Lead-Experience-Ride.html. This corrects a claim made elsewhere in this file: 04 does not have “no surface at all.” It has an experiment-grade harness already emitting lead_received, lead_assigned and assignment_escalated with server-side timestamps. What it lacks is a product surface, and the lead-economics instrumentation the stop actually needs.
Deployed to at least one pilot store and in daily use by the seat that owns the stop.
R12
A prototype gets mistaken for a capability. The roadmap assumes it is done, the benchmark is never taken, and the first real customer meets an unfinished surface. Attribution: process.
As the BDC rep or Outfitter, I consult on the customer’s actual question across at least two channels within the cadence floor, until two-way contact is established or the floor is exhausted and the lead is dispositioned.
Boundary only. Exit criteria exist; what the customer needs to understand here does not.
What must the first human response contain to earn a reply: an answer, a better question, or an appointment?
A cold customer can state, in their own words, the one thing this stop was meant to make clear.
P1P3F1
Surface defaults to inventory-first or form-first. Customer proceeds without comprehension, then stalls later at a stop that assumed it. Attribution: system.
Two-way exchange established — the customer replies. Observable. What the first response must contain to earn that reply is the open question.
Reply rate is reportable and attributable to response content, not just speed.
Speed is measured and content is not, so fast worthless replies score the same as slow useful ones.
No agent assigned. The eight-agent roster is fully specified — goal, bounded authority, tools, feedback loops — and none is mapped here. "The system handles this" is an intention.
A customer can complete this stop with no human present, or the system states plainly why a human is required.
P6
The stop runs on human memory. Same customer gets a different experience per rep, per store, per day. Attribution: process.
Role: BDC or Outfitter. Verb: consult — advise on a decision. Exits on two-way contact established.
A rep can see the customer’s authored playbook before responding.
Reps respond without context and ask what the customer already answered. Attribution: person.
Outfitter Statistics carries rep ownership of the connection event. [TBD-P3] Response Rates does not exist. Metrics: lead-to-connected conversion, response time, rep credited per touch.
Both a speed number and a quality number exist for the same touch.
Type A
We optimise for speed because it is the only thing measured.
Prototype only — not deployed. Surfaces that exist: the same internet-lead-to-appointment.build harness — its console is the only operator surface, emitting first_attempt_sent and connect_validated. The harness enforces the cadence floor — three attempts across at least two channels — independent of rep memory. But nothing surfaces first-response time to the person who owns it, and there is no disposition UI for a lead that never connects.
Deployed to at least one pilot store and in daily use by the seat that owns the stop.
R12
A prototype gets mistaken for a capability. The roadmap assumes it is done, the benchmark is never taken, and the first real customer meets an unfinished surface. Attribution: process.
As the BDC rep or Outfitter, I ask for the commitment and put a dated, timed, store-specific appointment on the calendar.
Boundary only. Exit criteria exist; what the customer needs to understand here does not.
Before they book: how long, who they meet, what they must decide, or what they are NOT committing to?
A cold customer can state, in their own words, the one thing this stop was meant to make clear.
P1P3F1
Surface defaults to inventory-first or form-first. Customer proceeds without comprehension, then stalls later at a stop that assumed it. Attribution: system.
Appointment booked with date, time and store. Observable and unambiguous. What makes booking feel low-risk is the open L1 question.
tour_scheduledA customer can book without speaking to anyone, and receives a confirmation naming who they will meet.
Booking requires a human, so it only happens during business hours. Attribution: system.
No agent assigned. The eight-agent roster is fully specified — goal, bounded authority, tools, feedback loops — and none is mapped here. "The system handles this" is an intention.
A customer can complete this stop with no human present, or the system states plainly why a human is required.
P6
The stop runs on human memory. Same customer gets a different experience per rep, per store, per day. Attribution: process.
Role: BDC or Outfitter. Verb: ask for commitment. Exits on an appointment with date, time and store.
The confirmation names a specific person, not "a team member."
The customer arrives expecting nobody in particular, and the greeting cannot honour a promise nobody made.
Outfitter Statistics — per-rep appt-set is implicit in their funnel. Metrics: appointments set. Set-to-show is measured at Gate A2.
Both booking paths are separately visible.
Type A
No operating surface found in the active project tree as of 2026-08-19. Only a concept board — “making it easy for them to schedule a visit.” No scheduling surface, no calendar binding, no confirmation artifact.
A named person can complete this stop end to end on a real surface, with no spreadsheet and no verbal workaround.
R12
The stop exists on paper only. L1–L4 can every one read green while the store still has nothing to use — which is precisely how a definition-complete system ships zero value. Attribution: build.
Showing up. The single most valuable unmeasured conversion in the chain — a booked appointment that becomes a physical arrival.
Show rate is reportable per store and per booking path.
No-shows are absorbed as noise. The most expensive leak in Segment 2 is invisible.
The metric is show rate, and the journey map states plainly that it is not surfaced as a discrete metric. Outfitter Statistics rolls effort into Tier %, which is a different number. Listed as an open question in the map itself.
Show rate appears on the board, separate from Tier %.
Type B
The gate between digital and physical — where the most money leaks — has no number.
As the receptionist or kiosk, I identify the customer, confirm the authored profile rather than re-asking it, and assign an outfitter within 30 seconds.
Boundary only. Exit criteria exist; what the customer needs to understand here does not.
What proves their pre-arrival work carried over — and what do we show a walk-in who has none?
A cold customer can state, in their own words, the one thing this stop was meant to make clear.
P1P3F1
Surface defaults to inventory-first or form-first. Customer proceeds without comprehension, then stalls later at a stop that assumed it. Attribution: system.
Identified and matched to their pre-arrival work. Observable as check-in complete with the playbook attached. The unresolved half is the walk-in with no history.
customer_nodeA returning customer sees their own work at check-in; a walk-in never sees an empty state that implies they are behind.
P5
Every arrival starts from zero. The authorship promise from 02 visibly breaks at the door.
No agent assigned. The eight-agent roster is fully specified — goal, bounded authority, tools, feedback loops — and none is mapped here. "The system handles this" is an intention.
A customer can complete this stop with no human present, or the system states plainly why a human is required.
P6
The stop runs on human memory. Same customer gets a different experience per rep, per store, per day. Attribution: process.
Role: Receptionist or kiosk, configurable. Verb: ask. Exits when the customer is physically present and identified. Screen is built — 4-step wizard plus Receptionist New Check-In.
02 requirement: read the authored profile, never re-ask itThe outfitter greets the customer already knowing their travel party and tow vehicle.
R33
Re-asking. The single most trust-destroying moment available, because it proves the digital work was theatre. Attribution: person.
Outfitter Statistics — rep credited. Metric: check-ins per day. Check-in to discovery conversion is not surfaced.
The first in-store transition has a rate.
Type A
We cannot tell whether arrivals become conversations.
Prototype only — not deployed. Surfaces that exist: Floor Workspace (4-step check-in wizard, Outfitter Status Cards), B1 · Reception, and reception-to-discovery-flow. The best-covered stop in the chain. It also carries the 02 requirement that the outfitter read the authored profile rather than re-ask it — which is a build behaviour, not a policy.
Deployed to at least one pilot store and in daily use by the seat that owns the stop.
R12
A prototype gets mistaken for a capability. The roadmap assumes it is done, the benchmark is never taken, and the first real customer meets an unfinished surface. Attribution: process.
As the Outfitter (or the customer self-serving), I extract what has not been articulated and turn it into a fit profile complete enough to select units against.
Boundary only. Exit criteria exist; what the customer needs to understand here does not.
Smallest set they must understand about their own use case before options mean anything — and which can they only learn by being asked?
A cold customer can state, in their own words, the one thing this stop was meant to make clear.
P1P3F1
Surface defaults to inventory-first or form-first. Customer proceeds without comprehension, then stalls later at a stop that assumed it. Attribution: system.
Fit profile complete enough to select units. Observable via fit_profile_completed. Inherits the deferred "RVs fit your life" message from 01.
09The system can state whether a profile is sufficient to run MCP, and why not if it is not.
P6
MCP runs on partial fit data and picks badly, which surfaces as a Line-Up problem three stops later.
No agent assigned. The eight-agent roster is fully specified — goal, bounded authority, tools, feedback loops — and none is mapped here. "The system handles this" is an intention.
A customer can complete this stop with no human present, or the system states plainly why a human is required.
P6
The stop runs on human memory. Same customer gets a different experience per rep, per store, per day. Attribution: process.
Role: Outfitter or customer self-serve — configurable pre-arrival or in-store. Verb: extract — pull out what the customer has not articulated and turn it into requirements.
An outfitter’s discovery session adds to the profile rather than recreating it.
R33
Discovery is re-run from scratch in store. Attribution: person.
Outfitter Statistics. Metric: discoveries completed. Qualification capture rate is not surfaced.
We can see which discovery fields go unasked.
Type B
A shallow discovery and a deep one count the same.
Prototype only — not deployed. Surfaces that exist: discovery-and-overview, discovery-to-mcp-flow, the Check-in to Bluesheet Accelerator project, and customer-mirror question screens Q1–Q5. The mirror screens exist as images, not as a built customer-side surface — so the requirement forbidding outfitter-private data on the mirror is currently enforced by nothing.
Deployed to at least one pilot store and in daily use by the seat that owns the stop.
R12
A prototype gets mistaken for a capability. The roadmap assumes it is done, the benchmark is never taken, and the first real customer meets an unfinished surface. Attribution: process.
As the Sales Manager or the system, I review the fit profile and return three units plus one wildcard to both the customer and the outfitter.
Boundary only. Exit criteria exist; what the customer needs to understand here does not.
R9 hides slot strategy. What is the customer-safe version of "why these three"?
A cold customer can state, in their own words, the one thing this stop was meant to make clear.
P1P3F1
Surface defaults to inventory-first or form-first. Customer proceeds without comprehension, then stalls later at a stop that assumed it. Attribution: system.
The customer opens a line-up they believe was built for them. Observable as line-up viewed. "Believe it was built for them" is the open R9-constrained question.
A customer can say why each unit is in their line-up without seeing slot vocabulary.
R9R2
The line-up reads as inventory push. The customer walks units they did not choose and gives polite verdicts.
No agent assigned. The eight-agent roster is fully specified — goal, bounded authority, tools, feedback loops — and none is mapped here. "The system handles this" is an intention.
A customer can complete this stop with no human present, or the system states plainly why a human is required.
P6
The stop runs on human memory. Same customer gets a different experience per rep, per store, per day. Attribution: process.
Role: Sales Mgr or the system — configurable manager-run or system-run. Verb: review / evaluate. Exits on 3 + 1 sent.
A store can run MCP without a manager present and get comparable picks.
R9R2R13
MCP only happens when a manager is free. The stop becomes a bottleneck. Attribution: person.
Outfitter Statistics. MCP completion rate is not surfaced. System-run mode has no measure at all.
Slot-win distribution is visible — the MCP pick-quality signal.
Type AR32
We cannot tell whether the manager’s picks are good, only that picks happened.
Prototype only — not deployed. Surfaces that exist: the densest set in the chain — Manager-MCP-Nexus-v4, MCP-Customer-Case-Fit, MCP-Pick-The-Winners, mcp-mobile, mcp-tesla-layout, pick-me-a-winner. Many candidates, no candidate of record. The system-run mode named in L3 is unbuilt in all of them.
Deployed to at least one pilot store and in daily use by the seat that owns the stop.
R12
A prototype gets mistaken for a capability. The roadmap assumes it is done, the benchmark is never taken, and the first real customer meets an unfinished surface. Attribution: process.
As the Outfitter, I walk the unit in person — never virtually, never automated — and capture per-zone reaction as it happens.
Boundary only. Exit criteria exist; what the customer needs to understand here does not.
What must they understand before the walk so it produces a real verdict rather than a polite one?
A cold customer can state, in their own words, the one thing this stop was meant to make clear.
P1P3F1
Surface defaults to inventory-first or form-first. Customer proceeds without comprehension, then stalls later at a stop that assumed it. Attribution: system.
A real verdict cast — at least one Favorite (R7). This is the only stop whose lever was already ruled: R7 exists precisely to stop a polite non-verdict being read as a decision. Also satisfies U1.
A tour cannot produce a winner without an affirmative verdict, and the fallback asks outright.
R7R29U1
Unit 1 gets presented as "The Winner" on zero feedback — the source’s own "most dangerous state."
No agent assigned. The eight-agent roster is fully specified — goal, bounded authority, tools, feedback loops — and none is mapped here. "The system handles this" is an intention.
A customer can complete this stop with no human present, or the system states plainly why a human is required.
P6
The stop runs on human memory. Same customer gets a different experience per rep, per store, per day. Attribution: process.
Role: Outfitter. Verb: present / demo. PHYSICAL ONLY — never virtual, never automated, not configurable. U1. Confirmed absolute 2026-08-19: the day-one direct-checkout customer still walked the unit.
No deal exists without a recorded in-person walk.
U1R4
Sales close without a walk, U1 becomes aspiration, and verdict data is fabricated.
Outfitter Statistics · Trade Overview — the trade walk-around often begins here. Metrics: demos delivered, trade-tool valuation creation rate.
rankSource distribution — the outfitter demonstration-effectiveness measure (R30, R32)A rising customer_selected rate reads as a demonstration-quality alarm.
Type AR30R32
Prototype only — not deployed. Surfaces that exist: bishs_rv_walk_through_prototype.html, with a walkaround area-model and intake taxonomy spec alongside it. U1 is physical-only and absolute, so the surface is a capture instrument and never a substitute for the walk. Per-zone capture and the customer-mirror engagement events the M/T/V grid asks for are not in the prototype.
Deployed to at least one pilot store and in daily use by the seat that owns the stop.
R12
A prototype gets mistaken for a capability. The roadmap assumes it is done, the benchmark is never taken, and the first real customer meets an unfinished surface. Attribution: process.
As the Outfitter, I present the built number with the trade and payoff inside it, deciding what shows unprompted and what shows on request.
Boundary only. Exit criteria exist; what the customer needs to understand here does not.
Which parts of how the number was built show unprompted, and which only on request?
A cold customer can state, in their own words, the one thing this stop was meant to make clear.
P1P3F1
Surface defaults to inventory-first or form-first. Customer proceeds without comprehension, then stalls later at a stop that assumed it. Attribution: system.
The customer engages the numbers — opens the breakdown, moves a slider, or accepts a matrix cell. Observable. Which parts show unprompted is the open R8-constrained question.
Engagement with the numbers is measured, not just presentation of them.
R8R38
We present and hope. No signal distinguishes a number seen from a number understood.
No agent assigned. The eight-agent roster is fully specified — goal, bounded authority, tools, feedback loops — and none is mapped here. "The system handles this" is an intention.
A customer can complete this stop with no human present, or the system states plainly why a human is required.
P6
The stop runs on human memory. Same customer gets a different experience per rep, per store, per day. Attribution: process.
Role: Sales Mgr / Outfitter. Verb: explain reasoning + market data. Exits when numbers are presented and the trade appraisal is attached. Screen built ~90%. Earliest stop where "buying here is different" can land.
Every number a manager presents has a market-data justification available on the same screen.
R8R38
Explanations vary by rep. Attribution: person.
DSR Summary (Fresh Sold) · Sales Mix (deal quality) · Daily Pacing ("Sold not Approved"). Metrics: Fresh Sold Deals, Fresh Sold Sales GP, Trade Rate, FE Margin distribution buckets.
estimatedWe can see how often customers are shown provisional numbers.
Type AR38
Prototype only — not deployed. Surfaces that exist: desking-prototype.html and Financing Term Selector (Nexus). Nothing built decides which parts of the number-build are shown unprompted versus on request — which is this stop’s open L1 question, and it resolves as a UI decision.
Deployed to at least one pilot store and in daily use by the seat that owns the stop.
R12
A prototype gets mistaken for a capability. The roadmap assumes it is done, the benchmark is never taken, and the first real customer meets an unfinished surface. Attribution: process.
As the Outfitter, I capture the commitment as a timestamped artefact with a customer-acknowledged response that is referenceable in a later desk dispute.
Boundary only. Exit criteria exist; what the customer needs to understand here does not.
At commitment, what must be unmistakable about what is fixed and what can still change?
A cold customer can state, in their own words, the one thing this stop was meant to make clear.
P1P3F1
Surface defaults to inventory-first or form-first. Customer proceeds without comprehension, then stalls later at a stop that assumed it. Attribution: system.
deal_created — commitment taken. The cleanest lever in the chain and already instrumented. Also Gate A3.
Commitment fires one event, and the customer can state what is now fixed.
R4
No agent assigned. The eight-agent roster is fully specified — goal, bounded authority, tools, feedback loops — and none is mapped here. "The system handles this" is an intention.
A customer can complete this stop with no human present, or the system states plainly why a human is required.
P6
The stop runs on human memory. Same customer gets a different experience per rep, per store, per day. Attribution: process.
Role: Sales Mgr. Verb: ask for commitment. Screen built.
R4
Commitment happens verbally and is recorded later, losing the timestamp.
Same set as Write-Up · Pipeline Overview "Lender Submission" stage. deal_state becomes "Deal"; Structurer cycle-time starts.
Type A
Prototype only — not deployed. Surfaces that exist: the desking-prototype covers deal structure. No commitment-capture surface exists. The canon requires a timestamped artifact carrying a customer-acknowledged response, referenceable in a later desk dispute — two fields and a customer-side tap that nothing implements.
Deployed to at least one pilot store and in daily use by the seat that owns the stop.
R12
A prototype gets mistaken for a capability. The roadmap assumes it is done, the benchmark is never taken, and the first real customer meets an unfinished surface. Attribution: process.
As the Outfitter, I consult on recovery rather than pressure, and I log the leak reason and originating stop so the loss is counted.
Boundary only. Exit criteria exist; what the customer needs to understand here does not.
What must they understand so leaving reads as a step forward rather than a failure?
A cold customer can state, in their own words, the one thing this stop was meant to make clear.
P1P3F1
Surface defaults to inventory-first or form-first. Customer proceeds without comprehension, then stalls later at a stop that assumed it. Attribution: system.
Not a forward conversion — a recovery lever. Observable as a second-look booking or a return ready to write. Type B: the measure is the size of the leak, and recovery is the counter-lever.
06 or 11Leak rate and recovery rate are both visible, by the stop they left from.
Type BR18
The largest population in the dealership leaves and is never counted or contacted.
No agent assigned. The eight-agent roster is fully specified — goal, bounded authority, tools, feedback loops — and none is mapped here. "The system handles this" is an intention.
A customer can complete this stop with no human present, or the system states plainly why a human is required.
P6
The stop runs on human memory. Same customer gets a different experience per rep, per store, per day. Attribution: process.
Role: Outfitter. Verb: consult — recovery, not pressure. Off-spine exit, not a forward stop.
An unanswered customer question becomes visible manager work.
R18Type B
Questions die unanswered and the nurture layer is decorative.
Zero measurement. No dashboard, no metric, no event. The largest population in any dealership, entirely uncounted.
07–11, recovery rateA leader can name which stop leaks most.
Type B
Every improvement to Segment 3 is unfalsifiable, because the loss it should reduce is not counted.
No operating surface found in the active project tree as of 2026-08-19. Chunk 3’s Visit Summary portal is the intended surface and is not built. The largest population in the dealership currently exits through no surface at all, which is the same reason its L4 is zero.
A named person can complete this stop end to end on a real surface, with no spreadsheet and no verbal workaround.
R12
The stop exists on paper only. L1–L4 can every one read green while the store still has nothing to use — which is precisely how a definition-complete system ships zero value. Attribution: build.
deal_created — the physical-to-deal-lifecycle crossing. Fully instrumented.
R4
Daily Pacing "Sold not Approved" amber flag · Delivery Dashboard "Create Delivery Order". Metrics: days stuck in Sold-not-Approved, approval success rate, Lender-Pending avg days vs rolling 30-day.
Type A
As the Finance Structurer, I take the handoff without the customer starting over.
Boundary only. Exit criteria exist; what the customer needs to understand here does not.
Do they need to see the finance handoff at all, or only to not be surprised?
A cold customer can state, in their own words, the one thing this stop was meant to make clear.
P1P3F1
Surface defaults to inventory-first or form-first. Customer proceeds without comprehension, then stalls later at a stop that assumed it. Attribution: system.
Deal packaged for submission. Internal event; the customer is not converting, the deal is progressing.
No agent assigned. The eight-agent roster is fully specified — goal, bounded authority, tools, feedback loops — and none is mapped here. "The system handles this" is an intention.
A customer can complete this stop with no human present, or the system states plainly why a human is required.
P6
The stop runs on human memory. Same customer gets a different experience per rep, per store, per day. Attribution: process.
Role: Structurer. Verb: process — back office.
R5
Daily Pacing "Approved not Scheduled" bucket · Pipeline Overview "Approved" stage. Metrics: Approved-not-Scheduled unit count, share of pipeline stuck pre-schedule.
Type A
No operating surface found in the active project tree as of 2026-08-19. Finance_CP12-20_Project and 04_Finance Funding both contain zero HTML/JSX.
A named person can complete this stop end to end on a real surface, with no spreadsheet and no verbal workaround.
R12
The stop exists on paper only. L1–L4 can every one read green while the store still has nothing to use — which is precisely how a definition-complete system ships zero value. Attribution: build.
As the Finance Structurer, I submit and record submit time, so approval rate and turnaround are both measurable.
Boundary only. Exit criteria exist; what the customer needs to understand here does not.
What must they understand about possible outcomes — including that the rate may move off the estimate?
A cold customer can state, in their own words, the one thing this stop was meant to make clear.
P1P3F1
Surface defaults to inventory-first or form-first. Customer proceeds without comprehension, then stalls later at a stop that assumed it. Attribution: system.
Submitted to lender(s). Observable.
R5
No agent assigned. The eight-agent roster is fully specified — goal, bounded authority, tools, feedback loops — and none is mapped here. "The system handles this" is an intention.
A customer can complete this stop with no human present, or the system states plainly why a human is required.
P6
The stop runs on human memory. Same customer gets a different experience per rep, per store, per day. Attribution: process.
Role: Finance Structurer — customer-invisible by definition (R5). Verb: process.
R5
Pipeline Overview "Lender Submission" → "Lender Pending" · Delivery Dashboard "Pending Finance Review". Metrics: bank submissions per day, Pending Finance Review avg hours.
Type A
No operating surface found in the active project tree as of 2026-08-19. No submission surface. Nothing shows the customer — or the seat — that a submission is in flight.
A named person can complete this stop end to end on a real surface, with no spreadsheet and no verbal workaround.
R12
The stop exists on paper only. L1–L4 can every one read green while the store still has nothing to use — which is precisely how a definition-complete system ships zero value. Attribution: build.
As the Finance Structurer or Producer, I explain any variance between approval and estimate before acceptance is requested, and it is recorded who explained it.
Boundary only. Exit criteria exist; what the customer needs to understand here does not.
When approval differs from the estimate, what is explained — and who says it, Structurer or Producer?
A cold customer can state, in their own words, the one thing this stop was meant to make clear.
P1P3F1
Surface defaults to inventory-first or form-first. Customer proceeds without comprehension, then stalls later at a stop that assumed it. Attribution: system.
Lender decision returned. External event on the lender’s clock, not ours.
A rate change never surprises the customer.
R38
No agent assigned. The eight-agent roster is fully specified — goal, bounded authority, tools, feedback loops — and none is mapped here. "The system handles this" is an intention.
A customer can complete this stop with no human present, or the system states plainly why a human is required.
P6
The stop runs on human memory. Same customer gets a different experience per rep, per store, per day. Attribution: process.
Deliberately none — external actor. The lender decides; nobody at Bish’s performs this stop. But the Submitted → Approved gap is where Structurer coaching directives live, so the transition is a human-performance measure even though the stop is not.
R5
Pipeline Overview "Approved" stage, separate from Lender Pending · Daily Pacing rollup. Metrics: approval count, Submitted → Approved conversion rate — the lender-fit decisioning gap.
Type A
No operating surface found in the active project tree as of 2026-08-19. No surface for the moment approval differs from the estimate, which is the moment this stop exists to handle.
A named person can complete this stop end to end on a real surface, with no spreadsheet and no verbal workaround.
R12
The stop exists on paper only. L1–L4 can every one read green while the store still has nothing to use — which is precisely how a definition-complete system ships zero value. Attribution: build.
As the Finance Producer, I present the menu, capture bought/not-bought per product, and file each signed document into the correct funding-packet slot automatically.
Boundary only. Exit criteria exist; what the customer needs to understand here does not.
Which documents get explained versus merely presented — and what is the standard for "explained"?
A cold customer can state, in their own words, the one thing this stop was meant to make clear.
P1P3F1
Surface defaults to inventory-first or form-first. Customer proceeds without comprehension, then stalls later at a stop that assumed it. Attribution: system.
All contracts signed — U2a. Observable and legally definitive.
Every document a customer signs was explained to a defined standard.
U2
No agent assigned. The eight-agent roster is fully specified — goal, bounded authority, tools, feedback loops — and none is mapped here. "The system handles this" is an intention.
A customer can complete this stop with no human present, or the system states plainly why a human is required.
P6
The stop runs on human memory. Same customer gets a different experience per rep, per store, per day. Attribution: process.
Role: Finance Producer — the customer-facing closer (R5). Verb: explain then ask for commitment.
The customer’s protection view and the desk’s view cannot drift — one catalog, sourced once.
R5R12
Menu presentation varies by producer. Attribution: person.
Pipeline Overview "Compliance/Doc Prep" → "Customer Signing" → "Stip Clearing" → "Contracted". Metrics: scheduled_unit_count, per-sub-step deal count and avg time, Producer avg days.
Type A
No operating surface found in the active project tree as of 2026-08-19. Doc Center logic is fully specified on paper — every-deal set, cash, finance, trades, final wet, funding packs, menu creation — and none of it is built. U2a depends on this stop.
A named person can complete this stop end to end on a real surface, with no spreadsheet and no verbal workaround.
R12
The stop exists on paper only. L1–L4 can every one read green while the store still has nothing to use — which is precisely how a definition-complete system ships zero value. Attribution: build.
As the Outfitter or delivery seat, I confirm the owner can perform each of the twelve checkpoints, not merely that they were shown.
Boundary only. Exit criteria exist; what the customer needs to understand here does not.
What must they be able to DO, not just have been shown, before the walkthrough is complete?
A cold customer can state, in their own words, the one thing this stop was meant to make clear.
P1P3F1
Surface defaults to inventory-first or form-first. Customer proceeds without comprehension, then stalls later at a stop that assumed it. Attribution: system.
Possession taken plus technician walkthrough — U3. Event exists (vw_deal.delivery_date). The surface does not.
No unit is released without a cert-matched tech and a completed walkthrough.
U3
The last mandatory step in the chain is invisible. A deal cannot be completed inside the system even though deals complete daily.
No agent assigned. The eight-agent roster is fully specified — goal, bounded authority, tools, feedback loops — and none is mapped here. "The system handles this" is an intention.
A customer can complete this stop with no human present, or the system states plainly why a human is required.
P6
The stop runs on human memory. Same customer gets a different experience per rep, per store, per day. Attribution: process.
Role: Certified Technician — flagged in the journey map as "a previously-unmodeled actor." Verb: present / demo + inspect. PHYSICAL ONLY.
Assignment fails loudly when no certified tech is available, rather than assigning anyone.
U3
Uncertified walkthroughs. OQS drops and nobody can attribute it. Attribution: person.
CIT Tracker is the primary dashboard — Contracts In Transit is exactly this state, delivered but not funded. Metrics: delivered_unit_count, Delivered Sales/Finance GP/NUFI, Avg Days to Deliver. The Orientation Quality Score in vw_survey is the customer’s own grade of this walkthrough.
Walkthrough quality is measurable per tech.
Type AR33
No operating surface found in the active project tree as of 2026-08-19. The sharpest case in the file. Every metric exists and the CIT Tracker covers it, while the stop has no screen and no chunk — and it carries U3 as a universal requirement. Fully measured, entirely unbuilt.
A named person can complete this stop end to end on a real surface, with no spreadsheet and no verbal workaround.
R12
The stop exists on paper only. L1–L4 can every one read green while the store still has nothing to use — which is precisely how a definition-complete system ships zero value. Attribution: build.
As the funding seat, I assemble the bank-specific packet and track CIT day count.
Currently scored none — back office. I do not think this is safe: they signed at 16 and walked at 17, then three stops pass in silence immediately after the two moments of maximum exposure.
Three stops pass in silence after signing and walking. Is silence correct, or do they need funding visibility?
Silence after signing reads as something going wrong. Inbound "what is happening" calls are the tell.
Funding package sent to bank. Internal event.
U2
No agent assigned. The eight-agent roster is fully specified — goal, bounded authority, tools, feedback loops — and none is mapped here. "The system handles this" is an intention.
A customer can complete this stop with no human present, or the system states plainly why a human is required.
P6
The stop runs on human memory. Same customer gets a different experience per rep, per store, per day. Attribution: process.
Role: back office. Verb: process — not customer-facing. Instrumentation only, no screen needed — the work already runs in EVO and the CIT tracker.
CIT Tracker — the active CIT inventory between Walked and Funded. Metrics: Bank-Pack inventory count (= CIT Current Count), Avg Bank-Pack days (= CIT Avg Age).
Type A
No operating surface found in the active project tree as of 2026-08-19. Segment 4 is the only fully measured block in the chain and contains no surfaces anywhere. Measurement without a surface is measuring a process people are running by hand.
A named person can complete this stop end to end on a real surface, with no spreadsheet and no verbal workaround.
R12
The stop exists on paper only. L1–L4 can every one read green while the store still has nothing to use — which is precisely how a definition-complete system ships zero value. Attribution: build.
As the funding seat, I record funded-in-period versus not-funded, and the interest cost of the gap.
Currently scored none.
Does "Bish’s has been funded" mean anything to them — or only that nothing more is owed?
Bank wire received — U2b. Bish’s is funded.
U2
No agent assigned. The eight-agent roster is fully specified — goal, bounded authority, tools, feedback loops — and none is mapped here. "The system handles this" is an intention.
A customer can complete this stop with no human present, or the system states plainly why a human is required.
P6
The stop runs on human memory. Same customer gets a different experience per rep, per store, per day. Attribution: process.
Role: Funder — not customer-facing. Verb: process. Instrumentation only.
CIT Tracker "Funded Count" and "Avg Days to Fund" KPIs. Metrics: funded_count, Avg Days to Fund = total_pipeline_days from finance_pipeline_health.
Type A
No operating surface found in the active project tree as of 2026-08-19. No funding surface. The CIT day count, and the interest cost of dollars sitting unfunded, are specified in the corpus and have nothing to display them.
A named person can complete this stop end to end on a real surface, with no spreadsheet and no verbal workaround.
R12
The stop exists on paper only. L1–L4 can every one read green while the store still has nothing to use — which is precisely how a definition-complete system ships zero value. Attribution: build.
As the seat that owns Finalized — currently unnamed — I close the deal record and trigger the ownership phase.
Currently scored none.
Is there a completion moment they should receive, or does the experience end at the walkthrough?
Books closed in EVO. Terminal event of the old 20-stop chain.
No agent assigned. The eight-agent roster is fully specified — goal, bounded authority, tools, feedback loops — and none is mapped here. "The system handles this" is an intention.
A customer can complete this stop with no human present, or the system states plainly why a human is required.
P6
The stop runs on human memory. Same customer gets a different experience per rep, per store, per day. Attribution: process.
Role: Accounting — not customer-facing. Verb: process. Open seat: who owns Finalized? Does not naturally fit the Funder role.
Finalized has a named owner.
The terminal stop of the deal lifecycle has no accountable person.
Daily Pacing "Finalized" bucket · DSR Summary Fresh Sold + MTD Sold. Metrics: finalized_unit_count, finalized_date set, FE + BE Margin + NUFI all credited.
Type A
No operating surface found in the active project tree as of 2026-08-19. No surface, and per L3 no seat either — “who owns Finalized” is still open. An unowned, unsurfaced stop is the definition of a silent failure point.
A named person can complete this stop end to end on a real surface, with no spreadsheet and no verbal workaround.
R12
The stop exists on paper only. L1–L4 can every one read green while the store still has nothing to use — which is precisely how a definition-complete system ships zero value. Attribution: build.
Finalized → first trip. New gate from D002. No lever defined, and the North Star sits on the far side of it.
The crossing is a named, observable event.
The chain ends at accounting. Nothing connects a closed deal to a camping trip.
No metric, no dashboard, no source. Added by D002 and never instrumented.
The gate has a rate.
Type A
The single most important conversion in the business is unmeasured.
As the Service Advisor, RVFix team or ACI — seat not yet assigned — I provide first-trip education and capture a first-trip completion event, which exists nowhere today.
Boundary only. Exit criteria exist; what the customer needs to understand here does not.
What must they know before the first trip that we assume they will figure out — and what failure are we most trying to prevent?
A cold customer can state, in their own words, the one thing this stop was meant to make clear.
P1P3F1
Surface defaults to inventory-first or form-first. Customer proceeds without comprehension, then stalls later at a stop that assumed it. Attribution: system.
First trip completed and feedback captured. This is the North Star, and no capture event exists in any built surface. Existing instruments measure first trip scheduled.
Plan Trip surface in Garage mode, or the portal blueprint’s first-trip planner deferred to V2A completed first trip fires one event, and feedback is attached to it.
The North Star is unmeasurable. Every claim about the mission is unfalsifiable.
No agent assigned. The eight-agent roster is fully specified — goal, bounded authority, tools, feedback loops — and none is mapped here. "The system handles this" is an intention.
A customer can complete this stop with no human present, or the system states plainly why a human is required.
P6
The stop runs on human memory. Same customer gets a different experience per rep, per store, per day. Attribution: process.
Verb defined: provide — education and resources. No role owns first trip. Candidates — Service Advisor, RVFix, ACI Insurance — exist as definitions with no screens. Same gap class as "who owns Finalized."
First trip has a named owner.
Nobody is accountable for the outcome the business defines as winning.
Missing — and this is the North Star. Ships labelled [NO SOURCE], not [BENCHMARK TBD]: better to show it empty than to omit it and measure only what is convenient.
Phase-6 completions are countable per store.
Type A
We report units sold because it is the only thing we can count.
No operating surface found in the active project tree as of 2026-08-19. The North Star has no surface. No capture event exists in any built artifact, which is why its L4 ships labelled [NO SOURCE]. Candidates named: the Plan Trip surface in Customer Garage mode, or the first-trip planner deferred to V2 in the portal blueprint.
A named person can complete this stop end to end on a real surface, with no spreadsheet and no verbal workaround.
R12
The stop exists on paper only. L1–L4 can every one read green while the store still has nothing to use — which is precisely how a definition-complete system ships zero value. Attribution: build.
As the Service Advisor or RVFix team, I consult and inspect, and the survey data that already flows finally has a consumer.
Boundary only. Exit criteria exist; what the customer needs to understand here does not.
How do they get help — and do they learn it at delivery, or when something breaks?
A cold customer can state, in their own words, the one thing this stop was meant to make clear.
P1P3F1
Surface defaults to inventory-first or form-first. Customer proceeds without comprehension, then stalls later at a stop that assumed it. Attribution: system.
First service appointment set. Observable. Whether the owner learns how to get help at delivery or at breakdown is the open question.
An owner can book service without calling.
Help is discovered at the worst possible moment, in a bad mood, by phone.
No agent assigned. The eight-agent roster is fully specified — goal, bounded authority, tools, feedback loops — and none is mapped here. "The system handles this" is an intention.
A customer can complete this stop with no human present, or the system states plainly why a human is required.
P6
The stop runs on human memory. Same customer gets a different experience per rep, per store, per day. Attribution: process.
Role named but unstaffed: Service Advisor / RVFix Team — definitions exist, no screens. Verb: consult + inspect.
The retention layer runs on phone calls.
Service cluster — 5 Pass-1 dashboards still unmapped. Metrics: first service appointment set rate; OQS / DQS / SQS / NPS; RVFix first-call resolution. vw_survey already flows and nothing consumes it — the chain ended before ownership.
vw_survey at this stopSurvey data reaches a dashboard.
Type A
We collect satisfaction data and never read it.
No operating surface found in the active project tree as of 2026-08-19. Service dashboards exist for reporting; no surface exists for the owner. vw_survey already flows and nothing consumes it — the wiring is there, the consumer is not.
A named person can complete this stop end to end on a real surface, with no spreadsheet and no verbal workaround.
R12
The stop exists on paper only. L1–L4 can every one read green while the store still has nothing to use — which is precisely how a definition-complete system ships zero value. Attribution: build.
As the system, I capture referral, advocacy and repurchase-readiness signal automatically, closing the cycle back to stop 01.
Boundary only. Exit criteria exist; what the customer needs to understand here does not.
What must be true before we ask — and what is the earliest honest moment to ask?
A cold customer can state, in their own words, the one thing this stop was meant to make clear.
P1P3F1
Surface defaults to inventory-first or form-first. Customer proceeds without comprehension, then stalls later at a stop that assumed it. Attribution: system.
Referral or repurchase signal captured. This is where the chain closes into a cycle — 23 loops to 01 — and the loop has no lever.
A referral is attributable to the customer who made it.
The cheapest lead source in the business is unmanaged.
No agent assigned. The eight-agent roster is fully specified — goal, bounded authority, tools, feedback loops — and none is mapped here. "The system handles this" is an intention.
A customer can complete this stop with no human present, or the system states plainly why a human is required.
P6
The stop runs on human memory. Same customer gets a different experience per rep, per store, per day. Attribution: process.
Deliberately none at MVP — automated capture. A human ask is optional and unassigned.
No metric, no dashboard, no event.
The loop back to 01 has a rate.
Type A
We cannot tell whether a good experience produces another customer.
No operating surface found in the active project tree as of 2026-08-19. Nothing built asks. This is where the chain closes into a cycle — 23 loops back to 01 — and the loop has neither a surface nor a metric.
A named person can complete this stop end to end on a real surface, with no spreadsheet and no verbal workaround.
R12
The stop exists on paper only. L1–L4 can every one read green while the store still has nothing to use — which is precisely how a definition-complete system ships zero value. Attribution: build.
One system that carries a customer from the first anonymous click to their first trip and back again — twenty-three stops with nothing dropped between them. Every stop knows what the customer needs to believe, what single act moves them forward, what the system decides on its own, which human is required and why, and what number proves it happened.
They stop being a lead and start being the author of their own buying playbook. What they build at the kitchen table survives the drive to the store — nobody re-asks it. They see how a number was constructed before they are asked to accept it. They leave able to operate what they bought. And their first trip works.
The outfitter stops guessing who to call and stops re-gathering what the customer already told us. Rotation is deterministic instead of argued. Every handoff carries a payload instead of a verbal summary. Nobody is measured on a number they cannot see, and no manager is asked to referee a dispute that the system should have recorded.
A chain that can be measured end to end, so improvement stops being anecdote. Cost per lead against value per lead. Benchmarks taken before change, not after. The bottom quartile moved up on purpose rather than by hope. And one definition of the process that survives turnover.
Written as the experience it will be. Each stop shows the moment in the customer’s words, the single observable act that delivers it, and who is present. Where the act is not yet named, it says so — this view is the vision, not a claim about what exists.
As someone not yet sure I can even do this, I spend two minutes answering one question and come away believing an RV fits how I actually camp — and nobody has asked who I am yet.
Completion of one micro-decision — the two-minute guide. Not dwell time, not pages viewed. A single self-guided act, because P3 says competence is proven by doing, never by being told.
No one. This is delivered by the system, on purpose.
As someone who just built something worth keeping, I save my buying playbook with a phone number and nothing more, and it is mine from any device, forever.
Saving the playbook — phone entered, magic link followed. Framed as preserving an asset they already built (P5), never as registering.
No one. This is delivered by the system, on purpose.
As a shopper with a trade and a tow vehicle, I get honest payment math instead of a teaser number — because I gave the four things that make the math real.
Camping Vision Worksheet plus tow specs submitted. Distinct from 02’s save: this is context, not identity, and it unlocks the four capabilities the math cannot run without.
No one. This is delivered by the system, on purpose.
As someone who asked a question, I get a named human, a clear next step, and the three things I gain by continuing: best fit, best market price, and every number before I ever drive to a store.
Not yet named. Until an observable act is defined here, this part of the vision cannot be built or measured.
First stop where a human enters the journey. Role: BDC or Outfitter, configurable per store. Verb: ask . Exits when the lead exists and is deduped to a customer.
As someone who replied, I get an answer to the question I actually asked — not a form letter, not a pitch, and not silence.
Two-way exchange established — the customer replies. Observable. What the first response must contain to earn that reply is the open question.
Role: BDC or Outfitter. Verb: consult — advise on a decision. Exits on two-way contact established.
As someone deciding whether to spend a Saturday on this, I know how long it takes, who I will meet, what I will be asked to decide, and what I am not committing to by showing up.
Appointment booked with date, time and store. Observable and unambiguous. What makes booking feel low-risk is the open L1 question.
Role: BDC or Outfitter. Verb: ask for commitment . Exits on an appointment with date, time and store.
As a customer who did the work at home, I walk in and they already know it. Nobody re-asks me a single thing I have already answered.
Identified and matched to their pre-arrival work. Observable as check-in complete with the playbook attached. The unresolved half is the walk-in with no history.
Role: Receptionist or kiosk, configurable. Verb: ask . Exits when the customer is physically present and identified. Screen is built — 4-step wizard plus Receptionist New Check-In.
As a customer, I am asked how I camp rather than what I want to buy — and the answers turn into a shortlist I recognise as mine.
Fit profile complete enough to select units. Observable via fit_profile_completed. Inherits the deferred "RVs fit your life" message from 01.
Role: Outfitter or customer self-serve — configurable pre-arrival or in-store. Verb: extract — pull out what the customer has not articulated and turn it into requirements.
As a customer, I am shown three units and one wildcard, and I understand why these three — without being shown the dealership’s internal math.
The customer opens a line-up they believe was built for them. Observable as line-up viewed. "Believe it was built for them" is the open R9-constrained question.
Role: Sales Mgr or the system — configurable manager-run or system-run. Verb: review / evaluate . Exits on 3 + 1 sent.
As a customer walking a unit, I record what I actually like and dislike as I go, so the verdict at the end is honest instead of polite.
A real verdict cast — at least one Favorite (R7). This is the only stop whose lever was already ruled: R7 exists precisely to stop a polite non-verdict being read as a decision. Also satisfies U1.
Role: Outfitter. Verb: present / demo . PHYSICAL ONLY — never virtual, never automated, not configurable. U1 . Confirmed absolute 2026-08-19: the day-one direct-checkout customer still walked the unit.
As a customer seeing a number for the first time, I can watch how it was built — MSRP, discount, Bish’s price, term, payment — with my trade and my payoff inside it.
The customer engages the numbers — opens the breakdown, moves a slider, or accepts a matrix cell. Observable. Which parts show unprompted is the open R8-constrained question.
Role: Sales Mgr / Outfitter. Verb: explain reasoning + market data . Exits when numbers are presented and the trade appraisal is attached. Screen built ~90%. Earliest stop where "buying here is different" can land.
As a customer saying yes, I know exactly what is now fixed and what can still move, and there is a record of it that neither side can quietly rewrite.
deal_created — commitment taken. The cleanest lever in the chain and already instrumented. Also Gate A3.
Role: Sales Mgr. Verb: ask for commitment . Screen built.
As someone leaving without buying, I leave with what we figured out together and a real reason to come back — not the feeling that I wasted a Saturday.
Not a forward conversion — a recovery lever. Observable as a second-look booking or a return ready to write. Type B: the measure is the size of the leak, and recovery is the counter-lever.
Role: Outfitter. Verb: consult — recovery, not pressure. Off-spine exit, not a forward stop.
As a customer who has committed, the handoff to finance is calm and I am not handed off to a stranger who starts over.
Deal packaged for submission. Internal event; the customer is not converting, the deal is progressing.
Role: Structurer. Verb: process — back office.
As a customer, I know what is being submitted on my behalf and what could come back different.
Submitted to lender(s). Observable.
Role: Finance Structurer — customer-invisible by definition (R5). Verb: process .
As a customer whose approval came back different from the estimate, someone explains why, in person, before I am asked to accept it.
Lender decision returned. External event on the lender’s clock, not ours.
No one. This is delivered by the system, on purpose.
As a customer signing, the documents that matter are explained and the rest are simply presented — and I can tell which is which.
All contracts signed — U2a. Observable and legally definitive.
Role: Finance Producer — the customer-facing closer (R5). Verb: explain then ask for commitment .
As a new owner, I can operate my RV before I drive it away. Not watch someone else operate it.
Possession taken plus technician walkthrough — U3. Event exists (vw_deal.delivery_date). The surface does not.
Role: Certified Technician — flagged in the journey map as "a previously-unmodeled actor." Verb: present / demo + inspect . PHYSICAL ONLY .
As a customer, the paperwork behind the scenes stays behind the scenes. Silence here is correct, so long as nothing more is owed of me.
Funding package sent to bank. Internal event.
Role: back office. Verb: process — not customer-facing. Instrumentation only, no screen needed — the work already runs in EVO and the CIT tracker.
As a customer, “funded” means nothing more is owed — and I am told that plainly rather than left wondering.
Bank wire received — U2b. Bish’s is funded.
Role: Funder — not customer-facing. Verb: process . Instrumentation only.
As a customer, there is a moment where this is finished and celebrated, rather than just trailing off after the walkthrough.
Books closed in EVO. Terminal event of the old 20-stop chain.
Role: Accounting — not customer-facing. Verb: process . Open seat: who owns Finalized? Does not naturally fit the Funder role.
As a first-time owner, my first trip works — because someone told me the things I would otherwise have learned the hard way, in the dark, in the rain.
Not yet named. Until an observable act is defined here, this part of the vision cannot be built or measured.
No seat assigned yet.
As an owner, I know how to get help before I need it, not the first time something breaks.
First service appointment set. Observable. Whether the owner learns how to get help at delivery or at breakdown is the open question.
No seat assigned yet.
As an owner who had a genuinely good experience, I am asked at the moment I would actually say yes.
Not yet named. Until an observable act is defined here, this part of the vision cannot be built or measured.
No one. This is delivered by the system, on purpose.
First trip completed. Not first trip scheduled — completed. It is the only measure that proves the whole chain worked, because a customer who camps successfully is the one who comes back and the one who refers. No capture event for it exists anywhere today, and it ships labelled [NO SOURCE] rather than quietly omitted.
Cost per lead set against value per lead. The conversion cascade — ups to write-ups to close to delivered — monetised through average front and back gross into expected contribution per outfitter. The lever is not exhortation: it is moving the bottom quartile up to the third, with the GP increase computed rather than hoped for.
A one-week benchmark before any workflow change, no exceptions. Server-side clocks, never client. A customer-acknowledged artefact at every commitment. And a customer-facing mirror that never shows internal data. These are not preferences; they are the reason a number from this system can be trusted.
This view is deliberately honest about its own gaps, because a roadmap that hides them stops being useful the first time someone checks. Three of twenty-one stops have an answered customer layer. Six conversion moments have no named act — and two of those six are the most valuable conversions in the business: getting a booked appointment to actually arrive, and getting a new owner through their first trip. Nothing in the chain is in production. Eleven stops have a prototype; thirteen have no surface at all. Switch to Tracker for the full state of every cell, and Dev mode for what exists to build from.
Everything a build team needs, per stop, in one place: the prototype of record, requirement documentation, event and data contracts, and QA criteria. Links are relative to this file and were verified against the project tree when this view was generated. Where a stop shows nothing, that is the finding — it is defined but not yet buildable, which is a different gap from unmeasured or unbuilt.
guide_started / guide_completed defined before build.lead_received (with source_system_ts preserved separately from server receipt), lead_assigned, assignment_escalated. Baton IN from M1 validated on lead_id, phone (E.164), source_token, source_channel, store_routing, timestamps and assigned_rep_id; name / email / vehicle-of-interest are soft-fail. Still missing for this stop: the lead-economics instrumentation — quality discriminator, cost, latency.first_attempt_sent, connect_validated, lead_deprioritized. Voice channel requires transcript_session_id on the baton out. Missing: a first-response-time surface, and a disposition UI for a lead that never connects (the finance handoff pattern — passed / terminal / store left — is the model).vw_survey already flows and nothing consumes it.01 and has neither surface nor metric.