Company-Building Blueprint

From a final decision memo to a company someone else can run

Take out your Week 17 final decision memo. It says which option the company should commit to, which assumptions the choice depends on, and what would make the company stop. The memo is a decision. A company is the work that follows the decision every day: who confirms the game time, who pays the performer, who checks each photo before it is published.

In the practice scenario, much of that daily work still lives in two people's heads. Mina knows which staff member at each team can really approve a change of date. Judy knows which sponsor needs a reminder before it pays.

In Week 7 you asked what would happen if Mina were sick on game day. The Week 8 organization chart named the translator as her backup, but that backup existed only on paper: she had not been trained, and the Week 9 audit had to ask for her training. Until a backup can really do the work, the honest answer is that someone would have to call Mina. This week's work removes that need.

A company can be inherited when another manager could run it tomorrow without calling the founder. The new manager might be a new hire, a partner, or Mina covering while Judy travels. To run the company, that person needs to find the reason for each rule, the person who owns it, the evidence behind it and what to do when something goes wrong.

The tool for this is a company blueprint: a written design of the whole company that another manager can use to run it. You drew a different kind of blueprint in Week 11. The service blueprint mapped one service, a game-day appearance, step by step, from what the fans see to the work behind the scenes. The company blueprint is larger. It covers the whole company, and the service blueprint becomes one of its parts.

The chapter continues the practice scenario, which is invented for learning. This week's question is: how can a final decision become a company that another manager can run, question and improve?

Where this week fits

In Week 17 you chose one option and defended it in a final decision memo. Week 18 places every artifact you have built since Week 1 into one company blueprint and tests whether a new manager could run it. It is the last step of the course, so the table ends with what you take beyond it.

One blueprint: six parts and an evidence appendix

The company blueprint in this course has six parts and an evidence appendix. You will use this same structure in the reading, the class practice and the assessment. Each part answers one question a new manager will ask.

Every part is written in the same way, as a short strip of four lines:

• Decision or rule: what the company does in this part of its work.
• Owner: who runs it, and who may pause it.
• Evidence and assumption: what supports the rule, with a pointer to the evidence appendix, and what is still assumed.
• Review trigger: the date or signal that makes someone look at the rule again.

The six parts join two ways of looking at a company. The first is the four questions from the course's reference book (Mitani): Target, Value, Capability and Revenue Model. They describe what the business is, and they appear in parts 2, 3 and 4.

The second is the four classic management functions, which describe how a business is managed. Planning chooses a direction and its limits (part 1). Organizing divides the work and the authority, and leading helps people do that work fairly and well (both in part 5). Controlling compares results with the plan and corrects the difference (part 6).

The same six parts work for every option, but each option puts its weight on different parts. For Option A · Expansion (expand cheerleading and sports-entertainment talent management), parts 4 and 5 carry the most weight, because cash timing and coordination are where Week 17 found the difficulty. For Option B · Boutique (a small cross-border talent and advisory service for a few clients), part 3 matters most: Judy's way of advising clients must be written down so that someone else can serve them. For Option C · Content-centered return (return to content, advertising, food service and personal brand), parts 1 and 3 matter most. The scope must say which lines of work the company runs, and each line needs its own routines and owner.

Where each of your artifacts goes

You do not start the blueprint from a blank page. Each artifact you built this term goes into one part. The table shows where.

Two checks come out of this table. First, if a part has no artifact, write "unknown" in it and name the person who will fill the gap. Second, look for contradictions between artifacts.

For example, the Week 5 goal tree said that new sponsored-video deals would wait during the test, but the Week 9 audit found that one had been accepted in test week 4. The blueprint must say which rule holds from now on: either new deals wait, or Judy may approve an exception in writing. A contradiction like this is useful, because it shows exactly where the plan and the real work drifted apart.

Worked example: the Revenue Model and cash strip

This example writes part 4 of a blueprint. It starts from Group 1's final decision memo in Week 17, which recommended Option A · Expansion, limited to three teams, with a deposit from each new team. A group whose memo chose Option C would write a different strip, but with the same four lines.

One word in the strip needs explaining first. To escalate a decision means to pass it up to a person with more authority, here Judy, because it is too close to a limit for the owner to decide alone.

The cash check is one calculation. Take cash today. Subtract the full NT$35,000 for every booked appearance whose costs the company has not yet paid, including the new one, because everyone is paid before the fee arrives. Add any deposits that will arrive before those games. The result decides what happens next:

• At least NT$130,000: the bookkeeper may accept the booking.
• NT$100,000 to below NT$130,000: the booking is escalated to Judy, because it falls below the Week 16 dashboard's action threshold. Judy decides before the booking is accepted.
• Below NT$100,000: the booking is declined or redesigned. Nobody may accept it.

Here is the check in one week. Cash is NT$150,000, the planning figure from Week 17, and nothing is booked yet. On Monday, a new team asks for an appearance and agrees to the deposit. The check is NT$150,000 − NT$35,000 + NT$30,000 = NT$145,000. That is at least NT$130,000, so the bookkeeper accepts.

On Wednesday, the home team asks for an appearance in the same month. It pays 30 days after the game and pays no deposit. The check now includes both bookings: NT$150,000 − 2 × NT$35,000 + NT$30,000 = NT$110,000. That is above the floor but inside the escalation range, so the request is escalated to Judy. She might accept it, or ask the home team to move the date until the first fee arrives.

Suppose Judy accepts. On Friday, another new team asks for an appearance but refuses the deposit. The check is NT$150,000 − 3 × NT$35,000 + NT$30,000 = NT$75,000, which is below the floor. The bookkeeper declines without calling Judy, and the refused deposit triggers a review of the strip.

In one week the strip produced all three outcomes. The bookkeeper decided two requests alone and knew exactly when to call Judy for the other.

Notice what makes this strip inheritable. The rule states its reason, which is to protect people the company already owes. It gives the owner a clear range for deciding alone. It also says which assumption is still untested, so the next manager knows what to watch.

What must travel with the document

A written strip is necessary, but it is rarely enough. Argote and Ingram (2000) describe a company's knowledge as held in its people, tasks and tools, so moving knowledge means rebuilding it in a new place. A document can carry the rule itself. The background, the reasons and the working relationships have to be rebuilt with the new person.

Szulanski (1996) studied how companies move a good practice from one unit to another and found three common obstacles. In plain words:

• The new person lacks the background to use it. The blueprint answers with a worked example, like the cash check above.
• Nobody can explain why the routine works. The blueprint answers by writing the reason next to each rule.
• The two people involved do not get along. The blueprint answers with a handoff meeting and a named person to ask.

Here is how this looks in the practice scenario. The service blueprint says, "Confirm the game time with the home team two days before." What it does not say is that only the home team's operations manager can approve a change. Mina knows this from experience. To transfer it, Mina writes the name of the role next to the step, explains why it matters, and runs the next confirmation together with the new person.

Zollo and Winter (2002) argue that companies learn most deliberately when people talk through their experience and write down what they learned, not only when they repeat a task. So each strip's review trigger should lead to a short meeting: what happened, which assumption failed and which line of the strip changes. Week 7 described a dynamic capability as the ability to renew what the company can do (Teece et al., 1997; Teece, 2007). The review trigger is the moment when that renewal happens: the company notices a change, decides how to respond and rewrites the strip.

Write rules that help people decide

Week 16 introduced Adler and Borys's (1996) distinction between enabling and coercive rules. An enabling rule helps a person understand the work, handle a problem and use judgment within a visible limit. The deposit rule is enabling: it states its reason, its range and the route to Judy. A coercive rule demands compliance and leaves no way to deal with the unusual. "Follow the schedule," with no way for the performer to report unsafe travel, is coercive.

A blueprint makes judgment visible by sorting its rules into three kinds.

Rules also need upkeep. Each strip carries a version date and the name of the person who may change it. Anyone who finds a rule that no longer works should be able to report it without having to prove loyalty first. If nobody can correct a rule, people quietly work around it. The blueprint then describes one company while the staff run another.

Keep the core, adapt the edge

Week 4 showed how to compare settings by mechanisms, such as who approves and which language controls. The blueprint uses that method when a routine moves to a new place or a new person. Kostova and Roth (2002) found that how fully a unit adopts a practice from head office depends both on local conditions and on the unit's relationship with head office.

So the blueprint separates the core from the adaptable edge. The core is the purpose or safeguard that must stay the same everywhere, such as written approval before an image is published. The adaptable edge is the way the core is carried out, which may change with the setting. For example, the performer might approve photos in Korean through a messaging app rather than by signing a printed form. An adaptation record states what changed, why, who approved it, and how the core is still protected.

Keep room to learn

Week 16 introduced March's (1991) distinction between exploitation, improving what already works, and exploration, trying what is new. A blueprint mostly records exploitation: the routines that deliver today's promise. It also needs a place for exploration.

Levinthal and March (1993) argue that organizations tend to learn most from recent, nearby and successful experience, and to overlook distant events and failures. The blueprint can correct for this in its review meetings. Each review asks at least one question about what the dashboard cannot see: a test that failed, a person whose voice is missing, or a signal too small to appear as a number.

Where AI can help, and this week's risk

An AI tool can turn your strips into a tidy handoff summary. The risk this week is that a summary keeps the rules and drops the reasons and the exceptions, which are exactly what a new manager needs. Check every summary against the strips, and keep the reason line. The NIST Generative AI Profile (Autio et al., 2024) recommends naming a person accountable for AI-assisted work; in a blueprint, that is the owner of each strip.

The inheritance test: a normal day and a difficult day

You can only know whether a blueprint can be inherited by testing it. Give it to someone who has never run the company, and ask them to handle two days using the blueprint alone. The example below continues the practice scenario.

A normal day. A team emails to confirm the game time and asks for two extra photos for its sponsor. The new manager finds the confirmation step in part 3 and the photo rule in part 6. Written approval is a non-negotiable safeguard, so she asks the performer before sending anything. Nothing here needs Judy, and the test passes because the manager could act and knew why.

A difficult day. In the same week, the performer says the home team has scheduled a promotional event that would leave her less than 12 hours of rest before the game. The home team's fee is also 9 days late. The new manager checks part 5, finds that the performer may refuse, and confirms the refusal to the home team; the performer liaison tells the performer in Korean. Part 4 tells the new manager that the late fee is a stop signal, so she pauses new bookings with that home team until the fee is paid.

How to handle the relationship with that home team changes a promise, so that question is passed up to Judy. The new manager records both items in the meeting decision log from Week 10 with the status "disputed", and updates them to "closed" when they are settled.

What the test found. The new manager could not find who at the home team approves schedule changes, because that contact was only in Mina's phone. Judy's staff revise part 5: they add a list of each partner's approver, owned by Mina and reviewed every season. Then they run the test again to check that the new manager can now find the contact.

The result of an inheritance test is a list of gaps, not a score. Common gaps are a rule with no reason, a person with responsibility but no authority, and a correction route that nobody can find. Each gap becomes a revision, and each revision is tested again with the people who use it.

Class practice: build and test your company blueprint

This practice is done without AI tools. Your teacher will tell you which worksheet to use and how to submit it. Bring your final decision memo and your earlier artifacts.

1. Place your artifacts. Use the table in "Where each of your artifacts goes". Write down any part that has no artifact, and one contradiction between two artifacts.
2. Write the core decision strip. Write the four lines of part 1: decision or rule, owner, evidence and assumption, and review trigger.
3. Write one more strip. Choose the part that your final decision memo depends on most, and write its four lines.
4. Run the inheritance test. Swap with a classmate. Your classmate acts as the new manager on a difficult day, such as a late payment, a refusal or a translation that changes the meaning of a promise. Write down one moment when they would have needed to call the founder, and one revision that removes that need.

Label any number you invent as invented. A strong blueprint lets your classmate act on the difficult day and explain why.

Looking back at the course

Before you finish, look back at your own work. Which artifact was hardest to connect to the others, and why? Which idea did you use in a later week without being sure you understood it? Go back to that week and reread one section. A manager who knows where their understanding is weak also knows what to check first.

The questions you practiced will work far beyond this case. A student club handing over to next year's committee can use the same six parts and the same normal-day and difficult-day test. The questions can be reused anywhere, but the answers must be found again, because each organization has its own payers, rules and people.

This is also where the course story ends. In Week 1, requests arrived at Judy's company faster than anyone could answer them. In Week 2, the company hired Mina as a bilingual coordinator, and she became the switchboard for every request. Over the term, you turned that flood into a decision memo, then into goals and tests, and finally into a company design. The case does not tell us which option Judy will choose, and your blueprint does not need to know, because it shows how any of the three options could be run by people other than the founder.

Key points: the end of the course

• A company can be inherited when another manager could run it tomorrow without calling the founder.
• The company blueprint covers the whole company. The Week 11 service blueprint is one piece of it, inside part 3.
• One structure holds everything: six parts and an evidence appendix, each part written as a strip of four lines. Every artifact from Weeks 1 to 17 has a place in it.
• Good rules state their reason and fall into three kinds: non-negotiable safeguards, discretion within a range, and matters to pass up to Judy.
• Keep the core the same, adapt the edge to the setting, and test the blueprint with a normal day and a difficult day.

The company blueprint is the last artifact of the course. Keep it with your final decision memo, and try the inheritance test on the next organization you join. A future manager should be able to find the reason for a rule, the person who owns it and the evidence behind it. If all of that still lives only in the founder's memory, the organization has inherited a story rather than a system.