What this document is
Northwind Services is an illustrative composite, not a client. Its numbers are constructed to show how the method works, not to report a measured result. Nothing described here has been validated in production. Where this document describes a future state, it is describing a design, not an achieved outcome.
The Blueprint describes Northwind as a designed system. This document is where people live in it. It takes each function in the Blueprint and says how it actually runs: what it decides, what it commits to, what rhythm it moves on, what information it has, what the machine does, what the human owns, which meetings it holds, when it escalates, how it is measured, how it fails, and what a real day looks like.
It is organized the way the Blueprint is, by the value engines under each pillar rather than by traditional departments, plus the Lead function above them and the service delivery layer beneath them. That organizing choice is deliberate. Departments are an artifact of how information used to move. Value engines are the units that actually produce something a customer pays for or that actually reduce a specific uncertainty. If you write the daily playbook by department, you re inscribe the old coordination structure into the new design and the redesign quietly fails. So the chapters below are engines, not org boxes. Metrics come straight from the pillar scorecards. Escalations run through the five decision filters. Meetings sit in the layer architecture. Nothing here is invented alongside the Blueprint; it is the Blueprint made operational.
Two fields in the template need the real company and not a designer's guess: current cadence, meaning how each function actually runs its rhythm today, and examples, meaning the real concrete day. In the original engagement these fields were marked as needing field capture rather than filled with fiction, because a manager who runs on an invented example runs into a wall. In this public version the same discipline applies with one adjustment. Where the shared case study gives Northwind a before picture, this document fills the field with a clearly labeled illustration drawn from that before picture and nothing else. Where the shared case study is silent, the field stays marked, and the mark is the point: it names where a real organization must capture its own current state before it can design a replacement. Everything else, the future design, is derivable from the Strategy and the Blueprint and is written in full.
The before picture, used throughout as the running illustration, is Northwind's path from closed won to revenue recognition. Observed as a constructed example: a weekly status collection where delivery leads assemble progress into a shared format; a summary produced from that collection for the leadership review; a leadership review that spends most of its time establishing what is true rather than deciding what to do; reconciliation between the commercial system of record and the delivery system of record, because the two disagree about scope, dates, and value; and a recurring gap between the moment a project slips and the moment the customer and the executive learn about it. The constructed figures attached to it: the status collection consumes 96 hours a month across all participants, and the lag from a project slipping to the executive learning of it is 9 days. Those numbers are constructed for illustration. They were not measured.
A reading note. Each function below follows the same template so a manager can find any field fast: Purpose, Primary decisions, Primary commitments, Cadence, Information, What the machine does, What the human owns, Meetings, Escalations, Metrics, Failure modes, and Example. A recurring paragraph labeled "What this means at Northwind." carries the concrete illustration for each function.
Function: Leadership (the Lead layer)
Purpose. Set and hold direction, and make the judgment calls the rest of the system cannot make for itself, especially the tradeoffs between pillars.
Primary decisions. Where the company is going; how to resolve conflicts between the three pillars; which changes enter the Transformation System and which are protected from it; the calls where the five filters point in different directions.
Primary commitments. The mission and values; the direction the scorecards are steering toward; the promise that accountability for consequential decisions stays with named humans.
Cadence. Future cadence, as designed: a regular direction and portfolio rhythm, less frequent than the operating layers, focused on judgment rather than status because status arrives already current.
What this means at Northwind. Current cadence, illustrated from the before picture: leadership meets on a review that runs off a summary assembled from the weekly status collection, and most of the session is spent establishing what is true rather than deciding what to do. A real organization would need to capture how its own leadership time actually splits between direction, decision, and status. That measurement has not been made here.
Information. The full pillar scorecard, live; the state of changes in flight in the Transformation System; the decisions record showing which filter carried past calls.
What the machine does. Keeps the scorecard and the change portfolio current so leadership never spends its time finding out where things stand; surfaces cross pillar conflicts early; drafts the option set for a decision, with the tradeoffs laid out against the filters.
What the human owns. The direction itself, the values, and every tradeoff where filters conflict. The machine can show that Leveraged Sales is about to overrun Transferrable Delivery capacity; only a human decides which one gives. At Northwind the named human is the chief executive, and the accountability is not delegable to the system that prepared the options.
Meetings. A direction and portfolio review that is a decision meeting, not a report. Everyone arrives current, so the whole session is judgment.
Escalations. This is where escalations land. A decision reaches Leadership when the filters genuinely conflict or when a change touches something Protected. Decisions that do not require that judgment should not arrive here, and if they do, the lower layers are underpowered.
Metrics. The whole scorecard, plus the long horizon measure the fifth filter implies: enterprise value and acquisition attractiveness. Leadership is the one function measured on all nine numbers and on the trajectory of the company as an asset.
Failure modes. Leadership drifts back into status and status reporting, spending its scarce attention finding out rather than deciding. Or it makes cross pillar tradeoffs implicitly, by letting one pillar quietly win, instead of deciding them in the open.
Example. This field is where a real organization must capture its own current state: a recent actual cross pillar tradeoff, how it was decided, and by whom. The shared case study does not supply one, so none is invented here.
Function: Marketing (Leveraged Sales engine)
Purpose. Generate leveraged demand, growth that does not consume effort in proportion, so the top of the pipeline fills from reputation and repeatable motion rather than pure outbound labor.
Primary decisions. Where to aim, which segments and which of the four service lines to lead with; which messages and offers to run; how to turn the differentiating line into door opening demand for the rest.
Primary commitments. A pipeline contribution target feeding Bookings; a brand position Northwind can defend with buyers in IT and operations at companies of 1,000 to 10,000 staff.
Cadence. Future cadence, as designed: continuous demand motion with a regular review of what is producing pipeline.
What this means at Northwind. Current cadence: this field is where a real organization must capture its own current state. The shared before picture covers the path from closed won to revenue recognition, not the demand motion upstream of it, so nothing is filled in here. Marking the gap is more useful than guessing at it.
Information. Which campaigns and content produce pipeline that actually converts, traced through Bookings to closed revenue, not just leads generated; live pipeline contribution.
What the machine does. Runs and measures the repeatable motion, traces demand to revenue so the company knows what actually works rather than what looks busy, drafts and personalizes content, and surfaces which plays are producing.
What the human owns. The positioning, the creative judgment, the strategic choice of where to aim, and the relationships and reputation that make demand leveraged rather than bought. The named human is the commercial leader, who is answerable for the position whether or not the machine drafted it.
Meetings. A demand review tied to the Leveraged Sales scorecard: what produced pipeline, what did not, what to change.
Escalations. A proposed campaign or position that fails a filter, most likely the customer filter or the can we actually deliver it filter, goes up. Marketing that generates demand delivery cannot fulfill is a cross pillar conflict for Leadership.
Metrics. Pipeline as the leading number, and downstream contribution to win rate and deal velocity. Marketing is judged on pipeline that converts, not activity.
Failure modes. Measuring motion instead of pipeline, so activity looks healthy while conversion does not. Or generating demand the delivery pillar cannot transferably fulfill, buying growth that costs margin.
Example. This field is where a real organization must capture its own current state: an actual instance of a differentiating capability opening a door for another line. None is invented here.
Function: Bookings (Leveraged Sales engine)
Purpose. Convert pipeline into booked, well shaped work that delivery can transferably fulfill, at a velocity and win rate that make growth leveraged.
Primary decisions. Which deals to pursue and which to walk from; how to scope and shape a deal so it is deliverable and profitable; pricing and terms within guardrails.
Primary commitments. Bookings that match delivery capacity and shape; a win rate and deal velocity that hit the scorecard; deals scoped so Transferrable Delivery can actually deliver them without heroics.
Cadence. Future cadence, as designed: continuous deal motion with a regular pipeline and forecast review.
What this means at Northwind. Current cadence, illustrated from the before picture at the point where it touches Bookings: the commercial system of record and the delivery system of record disagree about scope, dates, and value, so someone reconciles them by hand before anyone can say what a booked deal actually became. The rest of the current deal rhythm is where a real organization must capture its own current state.
Information. The live deal picture assembled without manual updating; each deal's fit against the filters and against current delivery capacity; historical patterns of which deals win and which deals turn into delivery pain.
What the machine does. Assembles the deal and pipeline state continuously, so sellers stop building status; surfaces which deals fit and which are dilutive based on real history; drafts proposals and scopes from playbooks; forecasts from real signal rather than optimism. Critically, it keeps the two systems of record agreeing continuously rather than producing a reconciliation event.
What the human owns. The relationship, the trust, the negotiation, and the judgment of whether a deal is real and right. Closing is commitment on both sides, and that is human. The named human is the account executive on the deal, and the commitment does not transfer to the system that scored it.
Meetings. A pipeline and forecast review tied to the Leveraged Sales scorecard, arriving current so it is about deciding which deals to push and which to drop, not about updating the numbers.
Escalations. A deal that fails a filter, that exceeds pricing guardrails, or that delivery flags as unfulfillable escalates. The classic one, booking work that cannot be staffed, is a cross pillar conflict that belongs to Leadership, surfaced early by the machine.
Metrics. Win rate and deal velocity as the core scorecard numbers, with pipeline as the input. The deeper measure is booking quality: how well booked work matches deliverable, profitable capacity.
Failure modes. Booking to hit a number rather than to fit delivery, which moves the pain downstream and shows up later as blown utilization and missed on time delivery. Or slow velocity because sellers spend their time assembling status instead of selling.
Example. This field is where a real organization must capture its own current state: a specific recent deal the model would have flagged as dilutive, and whether it was pursued anyway. None is invented here.
Function: Project Governance (Transferrable Delivery engine)
Purpose. Deliver the booked work on time, transferably, and profitably, so delivery is a repeatable system rather than a set of heroic individuals.
Primary decisions. How to staff and sequence the work; when a project is diverging and what to do; what gets escalated to the customer and when.
Primary commitments. On time delivery; utilization that protects margin without burning people; delivery that runs from playbooks so it can transfer and scale.
Cadence. Future cadence, as designed: continuous project telemetry with divergence caught as it happens, and a light regular review rather than a heavy weekly status assembly.
What this means at Northwind. This is the function where the current state pain concentrates, and it is the one place the shared before picture speaks in full detail. Illustrated, and constructed rather than measured: delivery leads assemble progress into a shared format on a weekly status collection; a summary is produced from that collection for the leadership review; the collection consumes 96 hours a month across all participants; and the lag from a project slipping to the executive learning of it runs 9 days. Those figures are constructed for this illustration. In a real engagement, this exact field is the one that most needs field capture, because the redesign is measured against it. If the before picture is guessed, the after picture proves nothing.
Information. Live project state from the work itself rather than from status reports; early divergence signals; capacity and utilization across the portfolio, live.
What the machine does. Maintains project state continuously so no one assembles status; detects divergence early and prepares the intervention options before a milestone slips; tracks utilization live; runs the delivery playbooks. This is the sharpest instance of separating finding out from deciding anywhere in the company. The machine does the entire finding out: it reads the work, assembles the current picture, notices the slope of a milestone before it crosses the line, and lays out two or three responses with their consequences. It does not decide.
What the human owns. The delivery judgment, the client conversation, the call on what a divergence means and how to respond, and the craft of the work itself. The named human is the project manager on the engagement. The project manager validates and decides; the machine watches and prepares. When the customer is told a date has moved, a person says it and is answerable for it.
Meetings. A delivery review tied to the Transferrable Delivery scorecard, and customer facing conversations that now happen before a slip rather than after, because divergence was seen early. Note what this design removes: the weekly status collection has no place in it, because its only product was a picture the machine now holds continuously.
Escalations. A divergence beyond a threshold, a customer issue, or a resource conflict escalates. A staffing conflict with a new booking is the cross pillar tradeoff surfaced to Leadership.
Metrics. On time delivery and utilization as the scorecard core, with customer satisfaction shared with Account Management. The deeper measure is transferability: how much delivery runs from playbook versus from a specific senior person.
Failure modes. Discovering divergence late, after the milestone slips and the customer is already unhappy, which is the exact failure the live telemetry exists to kill. Or optimizing utilization in ways that burn the people or hurt the renewal. A third failure is subtler: the machine prepares the picture, and the project manager stops exercising judgment on it and simply forwards what the machine produced. That is the separation collapsing back into fusion, with the accountability quietly unowned.
Example. Illustrated from the before picture: the current path from a slipping milestone to the executive learning of it runs 9 days at Northwind, with the customer typically learning at or after that point. That constructed figure is the number the future state is measured against. A real organization has to walk this path and time it. The illustration is not a substitute for that walk.
Function: Account Management (Transferrable Delivery engine)
Purpose. Keep customers successful, retained, and expanding, and move them along the service catalog from Assess to Implement to Manage to Govern and Advise, across service lines.
Primary decisions. Where each account is healthy or at risk; where the cross sell and expansion paths are; when to intervene on a relationship.
Primary commitments. Retention and expansion; customer satisfaction; the cross line journey that turns a single service into a managed relationship.
Cadence. Future cadence, as designed: continuous account health sensing with proactive intervention, and a regular account review focused on risk and expansion.
What this means at Northwind. Current cadence: this field is where a real organization must capture its own current state. The one thread the shared before picture supplies is the gap between a project slipping and the customer learning of it, which is an account health signal arriving late. How the rest of the account rhythm runs today is not known and is not invented here.
Information. Live account health from delivery signal, usage, and sentiment rather than from a periodic check in; the map of where each account sits in the catalog and what the next service is; satisfaction, live.
What the machine does. Senses account health continuously and flags risk and expansion signals early; maps each account against the catalog to surface the next logical service; prepares the account picture for every conversation.
What the human owns. The relationship and the trust, the judgment of what an account really needs, and the expansion conversation, which is a commitment the customer makes to a person. The named human is the account manager, who owns the intervention call whether or not the machine flagged it.
Meetings. An account review tied to satisfaction and expansion, arriving current so it is about deciding where to intervene and where to grow, not about assembling account status.
Escalations. An at risk account beyond a threshold, or an expansion that would strain delivery, escalates. Expansion that outruns transferable capacity is again a cross pillar tradeoff.
Metrics. Customer satisfaction as the shared scorecard number, plus retention and expansion. Account Management is judged on kept and grown customers, not on activity.
Failure modes. Finding out an account is unhappy at renewal instead of months earlier, when the live sensing should have flagged it. Or missing the obvious next service on the catalog because no one held the cross line view.
Example. This field is where a real organization must capture its own current state: a specific account that churned or nearly did, and how early the signals were actually available. None is invented here.
Function: Revenue (Profitability engine)
Purpose. Engineer the economics, so margin, collections, and cost per delivery are managed outcomes rather than end of quarter surprises.
Primary decisions. Where margin is leaking and why; where collections are slipping; how to bring delivery cost in line without harming the work.
Primary commitments. Margin targets; collections discipline; a cost per delivery that keeps the Profitability pillar healthy.
Cadence. Future cadence, as designed: live financial state with issues surfaced as they form, and a regular profitability review that decides rather than reconciles.
What this means at Northwind. Current cadence, illustrated from the before picture: the commercial system of record and the delivery system of record disagree about scope, dates, and value, so recognizing revenue on a closed won deal requires a manual reconciliation between them. That reconciliation is the visible tip of the information tax in this function. How much of the wider close and collections process is manual assembly is exactly what a real organization must capture, and it is not estimated here.
Information. Live margin, collections, and cost per delivery, reconciled from source rather than assembled monthly; early signals of a deal or project trending unprofitable.
What the machine does. Keeps the financial numbers live and reconciled, which is close to pure uncertainty reduction and is a strong candidate for the largest single manual assembly cost in a company of this shape; flags margin and collections issues early; prepares the profitability picture. The reconciliation between the two systems of record stops being an event and becomes a continuous state.
What the human owns. The judgment about what a financial signal means and what to do, the calls on cost and pricing discipline, and accountability for the economic decisions. The named human is the finance leader, who signs the number and answers for it.
Meetings. A profitability review tied to the Profitability scorecard, arriving current so it is about deciding on real margin and collections issues, not about closing the books in the room.
Escalations. A margin or collections breach beyond a threshold escalates, as does a structural cost problem that only pricing or a pillar tradeoff can fix.
Metrics. Margin, collections, and cost per delivery, the whole Profitability scorecard. This function owns those three numbers directly.
Failure modes. Discovering a margin problem at the close, weeks after it became true and after the money is spent, which the live reconciliation exists to prevent. Or spending the function's time producing numbers instead of acting on them.
Example. This field is where a real organization must capture its own current state: how long it currently takes from a project going unprofitable to anyone knowing. The shared before picture times the slip to executive awareness at a constructed 9 days but says nothing about the margin signal specifically, so no figure is invented for it.
Function: Service Delivery (the four service lines)
Purpose. Deliver the actual expert work, through Northwind's four lines: Managed Infrastructure, Managed Applications, Managed Support Services, and Advisory and Assessment. Each runs the common catalog of Assess, Implement, Manage, and Govern and Advise.
This layer is where the Operating Model most needs the delivery organization, because each line runs differently and the honest current state of each, what is systematized and what is heroic, has to come from the people who do the work. What can be stated now is the shared design; the per line specifics are the next round of work, and they are exactly the kind of thing the Transformation System takes one at a time.
What holds across all four lines, by design. The catalog shape is common, so a customer moves Assess to Implement to Manage to Govern and Advise along a known path in any line, which is what makes cross line expansion real. The machine carries the continuous sensing load, most obviously in the monitored and managed services where watching at scale is not a thing humans can do by hand, and increasingly in the telemetry under every engagement. The human carries the craft and the judgment: the architecture calls, the tradeoffs where a customer's stated priority and its actual constraint conflict, the client trust. Each line runs from playbooks that the memory layer holds, so delivery transfers and scales rather than depending on a specific senior person, which is the whole point of the Transferrable Delivery pillar and the acquisition filter.
Metrics. The Transferrable Delivery scorecard applies across the lines: utilization, satisfaction, on time delivery, with margin and cost per delivery shared into Profitability. Per line, the deeper measure is transferability, how much of the line runs from system versus from heroes.
The strategic roles, carried from the Blueprint. Managed Support Services is the recurring anchor, so its operating model is optimized for retention and continuous managed value. Advisory and Assessment is the differentiator and the door opener, so its operating model is partly a demand engine, feeding Marketing and Bookings. Managed Infrastructure is the foundation, so it is the platform the other lines assume. Managed Applications is where the customer's own business logic lives, valuable both as a service and as the closest read on what a customer is actually trying to do. These roles should shape how each line is run, not just how it is sold.
What needs field capture, per line. The real current cadence, the real handoffs, the real bottlenecks, and the real examples for each of the four lines. None of that is supplied by the shared case study and none is invented here. Each line is a candidate for the diagnostic and the staged loop, and each earns its full operating chapter when its current state is captured rather than assumed. This is the largest remaining gap in the document, and naming it is more honest than filling it.
The Strategy says why a company of this shape must change. The Blueprint draws the company it changes into. This Operating Model says how each function runs inside that design. The Transformation System is how an organization gets from the marked current states here to the future states described, one tested change at a time. The functions written in full are the ones the Blueprint made concrete. The current cadences, the examples, and the per line delivery detail are marked, not missing by accident: they are the field capture that turns a strong design into a real operating manual, and they are the fastest, most honest next step. Everything on the future side of this document is a design. It has not been run.