Cases from the MBA and the Vanguard · MBA module — when a brand's standard lives in its founder's head and the first big order tests it
Subtitle: When a brand's integrity lives in one person, scaling tests it before it tests the market.
The order was one the brand had been working toward for three years. A regional retailer wanted a collection of naturally dyed, eco-printed garments, ten times the brand's usual monthly volume, with a second order to follow if the first went well and possibly getting another retailer on board.
The founder had prepared for this. She had built the brand on three promises: natural dyes, natural eco fabrics and responsible sources, and fair pay for the women who made the clothes. She had invested heavily in her people with a small maker space that trains new artisans. Senior tailors went on professional tailoring courses. Several team members were sponsored through further study. On paper, she had a team ready to grow.
It did not hold. Natural Dye batches varied more than the retailer's buyers accepted. Garments passed through three hands before anyone noticed a colour problem. Deadlines slipped. To recover, the team asked to buy cheaper synthetic dye for the base colours, or to subcontract to a larger workshop in China with different labour practices. The founder said no to both, and the delivery went out late and was partly rejected. Two of her best-trained tailors left within months for better-paid jobs at a larger manufacturer. One of them said, "You taught me everything except how to decide."
The founder looked for what had gone wrong. She could not say the training had been wasted because the team could dye, cut, print, and finish to a high standard. But every hard question still travelled up to her: Is this shade close enough? Can we substitute this fabric? Is this deadline worth a compromise? In the small workshop she had answered those questions in passing, sometimes without noticing she was doing it. The standard was not written anywhere. It lived in her judgment, and it was still hers at ten times the volume.
Skills had been developed, but authority had not. The team had been trained to make the product but never allowed to own the trade-offs between the brand's values and the client's deadline. Under pressure, that gap decided the outcome. People either waited for her or, when she wasn't there.
She saw three options.
The first was to grow only as fast as she could personally check the work. Integrity would be protected and the brand would stay recognisable. But it would decline the largest opportunity it had ever had, and it would confirm that the brand could not exist without her.
The second was to hand real authority to her production leads now: a written quality standard, a clear list of what may never be compromised, and the right to delay or refuse an order that would break it. Some decisions would be ones she would not have made, and at least one client might be lost. But the team would grow into ownership, and the standard would start to live somewhere other than her head.
The third was to bring in an experienced operations manager from conventional fashion manufacturing to build systems quickly. Delivery would improve fast. But that person might treat the dye rules and the pay structure as inefficiencies to remove, and the team she had invested in would read the appointment as a verdict on them.
There was also a personal question. Was she holding on to quality control because the brand's integrity required it, or because the brand was her, and letting others decide meant watching it become something slightly different?
She had begun writing SOPs and asking her leads to bring a proposal instead of a question. But every urgent order pulled her back into approving things. The retailers expected a steady scale-up, and the team expected her to have the final word. She knew that if she waited until the team felt ready, the second order would go to a competitor. If she pushed too early, the promises the brand was built on could be broken by the people she was trying to empower.
Before the next order is accepted, she has to decide which decisions leave her hands, which values are non-negotiable, and how much short-term risk to the business she will accept so that the brand's principles survive without her in the room.
Before any discussion, ask everyone to pick one of the three options: cap growth to what she can personally check, hand over real authority now, or bring in an experienced operations manager from outside. Ask first someone who has scaled a craft or values-led business. They usually know which standards are truly essential to the brand and which are only habits of the founder. Then ask someone from a conventional manufacturing or retail background what a buyer actually tolerates in colour variation and lead time, and whether the brand's promise can survive that.
The fact most likely to change the room's answer is what the team asked for when the deadline pressed. They did not ask to abandon the brand. They asked permission to compromise it, because no one had told them where the line was or who could hold it. If the room decides early that the team was not capable, take them back to that paragraph. If the discussion gets too abstract, ask one practical question: what is the one decision about dye, fabric, or pay that can be handed to a production lead next Monday, what are its limits, and what outcome must never happen? This case is a composite drawn from experiences in sustainable fashion. Names, locations, figures and organisational details have been generalised to preserve confidentiality.
The brand has no such system today. But imagine that before the next order she could give the team an AI operating agent built from her quality records, dye recipes, supplier rules, pay principles and the reasoning behind her past decisions. It could tell a production lead whether a shade is within standard, flag a substitution that breaks a sourcing rule, recommend a certified producer in China who meets the standard and show what happened the last time a similar deadline conflict came up. The brand's integrity would sit in something the whole team can consult, not in one person's head.
It could also deepen the dependence. If the agent learns only from her judgments, it will apply her taste and her risk tolerance at scale, and the team will wait for the agent as they once waited for her, without becoming any more able to decide. There is a second risk: an agent that knows the standard but not what a missed order costs the business can protect the principles and sink the brand. So the real question is whether the agent should enforce her standard or help her team understand it well enough to defend it themselves. And when a tailor follows a recommendation built on the founder's past choices, whose decision is it?
She built a brand people believe in. She has not yet shown that its principles survive and scale when she isn't the one holding them.
A live case: every round can be improved, and the author's feedback is the next one.
The same three questions, answered twice: first without the mentor's corpus, then from it — the Command Layer, training under conditions, the invisible wins, his decision doctrine and the task-to-agent manual.
Without the mentor's corpus
1. The second. The team did not ask to abandon the brand; it asked permission to compromise it. People who do not care about a promise do not ask before breaking it. What they lacked was its price. Patrick Lencioni warned in Harvard Business Review in 2002 that real values, properly practised, "inflict pain": they limit a company's freedom and constrain what its people may do. The founder paid that price for years, quietly and alone. Her team saw what the promises cost under a deadline, never what breaking them would cost. The second gap is who holds the cord. Toyota traces jidoka, its principle of stopping for quality, to Sakichi Toyoda's looms; his 1896 power loom already stopped itself when a weft thread broke. Today an operator who sees a problem on a Toyota line can stop it by pulling a cord. The right to stop sits with the person who sees the problem. In this workshop, only the founder could stop the line or let it run. So write each promise down with its price and its owner:
Ask yourself: which of my promises has its price and its owner written next to it?
2. Not wasted, but half-built. She invested in skills every employer values, and in nothing only her brand could give. Economists have drawn that line since Gary Becker's Human Capital (1964). General training raises a worker's value everywhere, so in a competitive market employers will not pay for it; specific training cannot be carried to another firm. In 2006 Anke Kessler and Christoph Lülfesmann showed in the Economic Journal that the two work together: the possibility of firm-specific training is what makes an employer willing to invest in general skills. Her tailoring courses were general. The specific part, how this brand judges a shade, a fabric or a deadline, was never taught, because it lived in her head. Evidence on turnover points the same way. A two-wave study of 588 employees found that, of six development activities, only promotion raised the risk of leaving, by improving outside prospects, while skill utilisation, the chance to use what one has learned, directly kept people. "You taught me everything except how to decide" describes skills that were never allowed to be used. So training for a values-led brand should add four things:
Ask yourself: what can my people learn and do here that a bigger manufacturer cannot pay them for?
3. Set the number before the order, and spend it on the team's decisions, never on the principles. Two tools help. Saras Sarasvathy, who studied how expert entrepreneurs decide, found that they start from affordable loss rather than expected return: they settle first what they can afford to lose, then act within it. Jeff Bezos's 2015 letter to Amazon's shareholders separates one-way doors, irreversible decisions to be made slowly and carefully, from two-way doors, which should be made quickly by people with good judgment. The brand's three promises are one-way doors: one synthetic base colour or one unaudited subcontractor breaks what the brand sells, and cannot be taken back. A shade inside an agreed band, a fabric swap from a certified list or a renegotiated date are two-way doors, and they belong to the production leads now. The affordable loss is the price of their learning on those doors: a late delivery, a smaller second order, perhaps one lost client, which option two already accepts. The principles set its ceiling. If lost orders put the wage bill at risk, the brand breaks its third promise while defending the first. Part of the loss can be avoided before the order: agree the shade band and the lead time with the buyer in writing. Ask yourself: what would I have to lose before I took a decision back, and does my team know that number?
Sources: P. Lencioni, "Make Your Values Mean Something", Harvard Business Review, July 2002; Toyota Motor Corporation, "Toyota Production System" (jidoka and the 1896 Toyoda power loom); G. S. Becker, Human Capital, 1964; A. S. Kessler and C. Lülfesmann, "The Theory of Human Capital Revisited: On the Interaction of General and Specific Investments", Economic Journal 116(514), 2006; J. Nelissen, A. Forrier and M. Verbruggen, "Employee development and voluntary turnover: testing the employability paradox", Human Resource Management Journal 27(1), 2017; S. D. Sarasvathy, "Causation and Effectuation", Academy of Management Review 26(2), 2001 (the affordable-loss principle); J. Bezos, letter to Amazon.com shareholders, 2015.
From the mentor's corpus
1. The second, and the mentor's Command Layer chapter describes this moment before it happened. A decision is made at the centre, conditions at the edge change, and the orders no longer fit. The subordinate then has three options: wait for new orders and lose the window, follow the old orders and produce the wrong action, or act on her own judgment with no cover. The chapter calls them catastrophe by inertia, catastrophe by compliance, and catastrophe for the subordinate if she was never given the authority to act. This team lived the first two. It waited for her, then followed the old orders into a late, partly rejected delivery. The third it avoided, rightly, because it had no cover, so it asked. That is not a failure to understand the values. The mentor's dissertation Recruiting Intelligence adds that when decentralised action fails, the leader cannot easily deny responsibility, because the failure often reveals a defect in intent design, information architecture or training. The chapter's remedy is intent written before orders, answering three questions in a form a subordinate can hold in her head under stress. For this order:
Ask yourself: could my production lead say what must be protected, the night before a shipment, without calling me?
2. Not wasted, but trained for the wrong conditions. The mentor's second volume states the rule in one line: training produces what training conditions for. When the conditions diverge, it produces a certificate and a problem. Her team trained on courses and in a small workshop, and met its first real test under a buyer's clock. The chapter on training under conditions lists the behaviour no course certifies, and each item has a place here:
None of it appears by declaration. It is built through rehearsal and after-action review, with losses by design. The same volume names what the departing tailor described. When the language of delegation sits on top of centralised execution, the subordinate "discovers, over time, that her own judgment was never actually authorised", and initiative atrophies. The first volume explains why the best leave first: the most capable people have options, and they choose environments where they have autonomy. Its test is the right measure of this investment: how many people can execute on your intent without direct oversight? Ask yourself: when did my team last rehearse a hard trade-off with me out of the room?
3. Accept a loss of margin, never of quality, and fix its size before the order. The mentor made this choice himself. In 2008, as enrolment collapsed across the region, COTRUGLI could cut quality and stay profitable, or keep quality and accept lower margins for a few years. It kept quality. Three difficult years followed, and by 2012 the school was growing while competitors closed. The first volume calls this an invisible win: a crisis prevented, which no quarterly report captured. Two rules from the mentor's work set the terms of the loss. His decision doctrine holds that roughly 80% confidence can be enough for a reversible decision. So a shade inside the band, a certified fabric substitute or a renegotiated date belong to the production leads now, and the loss budget must cover their mistakes too. The Command Layer adds that trust is built through consistent backing of subordinates who have acted within intent and taken losses doing so. A founder who takes the decision back after the first loss has not delegated. The first volume says where that leads: a leader who cannot accept this risk cannot practise mission command and reverts to micromanagement. That explains why her SOPs have not held: every urgent order still pulls her back into approving. Ask yourself: which loss would I back in front of my team, if my production lead made it inside my intent?
On the NEO Turn. The mentor's task-to-agent manual answers the first question in one line: if the data lives in someone's head, the agent cannot reach it. So the agent's first effect comes before it runs. To build it, the founder must write down the standard she has never written, and then her team can read it too. What it does next depends on the purpose it is given. The mentor calls AI a teleology magnifier: it amplifies the purpose behind its deployment. Built to enforce her taste, it scales the bottleneck; built to develop judgment, it scales that. The manual's filter draws the line. Checking a shade or a supplier against written rules is calculator work: the agent processes, a person reviews. Trading a value against a deadline is craft, which the protocol never automates. The charter must also say what each order has to produce, or the agent will protect the principles and sink the brand. Whose decision is it? The Command Layer answers: the tailor is responsible for the action, the founder for the intent, the boundaries and the authorisation. One kill indicator belongs in the charter: if the questions that travel up do not fall month by month, the agent is teaching the team to wait.
Sources: Vanguard Leadership, vol. 2, the Command Layer (the failure mode the doctrine prevents, intent before orders, accountability and trust under delegation), "Training Under Conditions", and the failure modes of mission command in the full manuscript; the mentor's dissertation Recruiting Intelligence, chapter 4; Vanguard Leadership, vol. 1, §2.4 (the orchestration test), the principles of mission command (p. 182) and "The Invisible Wins" (p. 386); the mentor's decision doctrine of 22 August 2026 (§6); the Vanguard Task-to-Agent Mapping Protocol, academic edition (the Agent Definition Card, the Craft filter, the kill indicator); the mentor's glossary, the teleology magnifier.
The most honest sentence in this case is the tailor's: "You taught me everything except how to decide." The mentor's work studies a company that learned to answer it at scale: Haier. Two moments from its story fit this brand.
The first year: make the standard visible. In 1985, in his first year running a near-bankrupt refrigerator factory in Qingdao, Zhang Ruimin had 76 defective refrigerators destroyed with sledgehammers on the factory floor. In the mentor's reading, the act shattered the status quo and set a culture of quality consciousness that defined Haier for decades. This brand needs its own version of that moment, without destroying anything. Before the next order:
The later reform: move the decisions to the work. Haier went on to dissolve its hierarchy into thousands of micro-enterprises of about ten people, each accountable directly to the users it serves. A brand this size needs three, not thousands: dyeing, printing and finishing. Each cell is led by a production lead who owns its decisions inside the written standard and answers for its outcome:
The founder keeps the three promises and the relationship with the buyer. When a lead takes a loss inside that intent, she backs the lead in front of the team.
The share. The two tailors left for more money. Fair pay was already a promise; ownership adds a share. When an order ships on time and in band, the cell that made it shares in the margin. A larger manufacturer can match a wage. It is harder for it to match the right to decide and a share in what those decisions earn.
The next order is the test. If it ships late but in band, with decisions the founder would not have made, the brand has started to live outside her head. Ask yourself: which of my standards could my team defend in front of a buyer tomorrow, without me in the room?
Sources: the mentor's case study "Haier's Rendanheyi Model: Business Case Analysis of Leadership Innovation for the NEO Era"; Vanguard Leadership, vol. 2 (the Haier chapter); the mentor's article on NEO Cotruglian philosophy for the International Leadership Journal (Haier's micro-enterprises); "Zhang Ruimin's Haier Power", TIME, April 2014, republished by Haier Group.
She is a small entrepreneur, and she must not be afraid to decide. If she has chosen the path of natural dyes and everything that goes with it, this is the way.
She cannot afford an experienced manager, but she can work alone all day and learn how production is organised.
The big order was an opportunity of the kind that does not come often, and she has to meet it ready. Mistakes happen; let her learn from this one.
A question for the table, a disagreement, what you would have done. The case lead reads every comment; the ones the table takes up enter the chapter as questions from the room, with your name.