The first board pack named the risk nobody wanted to talk about. Now the board wants a plan to address it. Diane runs a workshop that makes everyone uncomfortable – and Dave puts something on paper for the first time in his life.
The risk register from the board pack listed key person dependency as GreenBox’s number one risk. Patricia Osei, in her first formal feedback to Maya, was characteristically precise: “Your company has excellent unit economics, a strong brand, a defensible moat in farm relationships, and a valuation discount of 20% because three people are holding everything together. Fix that, and the next conversation with investors is very different.”
The 20% discount isn’t a metaphor. It appeared in Diane’s valuation work during the What Are We Worth exercise. GreenBox’s revenue, growth rate, and retention metrics suggest a pre-money valuation around $44M on a standard revenue multiple. But investors apply discounts for structural risks. Key person dependency – the risk that the business would be materially harmed by any individual’s departure – is the biggest one. At 20%, that discount reduces the valuation from $44M to roughly $35M.
Nine million dollars. That’s what it costs GreenBox that Maya, Tom, and Dave are irreplaceable.
The bus factor exercise
Diane runs the workshop on a Wednesday afternoon in the Fremantle office. Present: Maya, Tom, Sam, Jas, Charlotte (video), Marcus, and Patricia (observing). Lee isn’t there – he’s in Margaret River, and this is an operational exercise, not a coaching session.
Diane has done this exercise with a dozen companies. She’s direct about it. “This isn’t comfortable. I’m going to ask you to imagine that each of your colleagues disappeared tomorrow. Not fired, not on leave – gone. And then I’m going to ask: what breaks?”
She draws a simple table on the whiteboard.
| Person | What they hold | Where it lives | Backup |
|---|---|---|---|
“For each critical function in the company, I want to know: who’s the only person who can do it? Where does the knowledge live – in a system, in a document, or in their head? And who’s the backup?”
The room is quiet. Tom shifts in his chair. This is the kind of exercise he’d normally dismiss as “process theatre” – but the $9M valuation discount has a way of focusing attention.
Maya
They start with Maya. Diane asks the team to list everything that would break if Maya disappeared tomorrow.
The list grows longer than anyone expects.
Customer relationships. Maya personally knows forty or fifty of the early subscribers. Mrs Patterson – the Stirling Highway subscriber who hates beetroot and has been loyal since week two – has Maya’s personal mobile number. Six of the farm partners, including Dave, have relationships with Maya that predate GreenBox. When a new farm partner signs on, Maya does the first visit personally. She knows which farms can flex on volume, which ones need advance notice, which ones have family situations that affect delivery schedules.
Where does this live? Mostly in Maya’s head. Some of it is in the CRM. The farm relationships are partially documented through the work they did in Decision Tables – Maya’s substitution knowledge was codified into rules that any team member can follow. But the relationship context – the trust, the personal history, the judgement about when to push and when to back off – isn’t in any system.
Brand voice. Maya writes the customer emails. She approves every piece of marketing. She reviews the social media posts. The GreenBox voice is Maya’s voice – warm, personal, slightly self-deprecating, always rooted in the farm connection. Jas and Sam can approximate it, but they check with Maya before anything goes out.
Strategic vision. The board deck, the growth plan, the city expansion strategy, the response to the Hartland Group acquisition – these come from Maya. Charlotte provides the scaling framework. Diane provides the business strategy. But the vision of what GreenBox is and where it should go is Maya’s.
Decision authority. Despite the management layer that Tom, Sam, and Jas now provide, Maya is still the final decision-maker on anything significant. Pricing changes, new city launches, farm partner agreements, hiring above a certain level – they all go through Maya.
Tom says it bluntly: “If Maya disappeared, we could keep the lights on for three months. After that, the farm relationships would start fraying, the brand voice would drift, and nobody would know what strategic decision to make next.”
Maya stares at the whiteboard. She’s not offended. She’s frightened.
Tom
Tom’s bus factor assessment is different but equally uncomfortable.
Architecture decisions. Tom designed the core GreenBox platform. The bounded contexts from the DDD work are documented. The ADRs capture the reasoning behind major technical choices. But the mental model of how everything connects – which service talks to which, where the performance bottlenecks are, why certain decisions were made early on that now constrain the system – lives in Tom’s head. The ADRs capture individual decisions. Tom holds the narrative that ties them together.
Deployment knowledge. Tom knows how to deploy. Priya knows how to deploy. Kai has done it a few times. But the deployment process is manual, and the edge cases – what to do when the staging environment is inconsistent, how to handle a failed migration, where the database backups actually are – are Tom’s tribal knowledge.
Codebase intuition. When something breaks at 8pm, Tom can usually identify the cause within twenty minutes. Not because he’s the best debugger – because he wrote or reviewed most of the critical code paths. That institutional knowledge accelerates everything. Without it, debugging takes hours instead of minutes.
Priya, on video from Melbourne, adds: “Tom also does something nobody talks about. He’s the person who spots when two squads are about to build the same thing. He sees the overlaps before they become conflicts. We don’t have a system for that. We have Tom.”
Tom looks at the whiteboard and says nothing for a long time. Then: “I always thought being indispensable was a good thing.”
Charlotte, on the call, says quietly: “It is. Until it isn’t.”
Dave
Dave isn’t in the room. He’s in Margaret River, doing what he does every Wednesday – working the farm. But his name goes on the whiteboard.
Farm relationships. Dave Morrison’s handshake with Maya in a farmers’ market three years ago is the foundation of GreenBox’s supply chain. His farm was the first partner. His network introduced Rachel and the other Margaret River farmers. His son Ben now manages the day-to-day relationship, but the trust is Dave’s. The other farmers partnered with GreenBox because Dave did. If Dave left – retired, got sick, decided he was done – would the farms stay?
Supply chain credibility. Dave’s presence gives GreenBox legitimacy with new farm partners. When GreenBox expands to a new region, the pitch is: “We’re a produce-box company with a three-year partnership with farms in Margaret River. Here’s Dave Morrison’s number – call him.” That reference is worth more than any marketing deck.
Domain knowledge. Dave knows farming. He knows what’s in season, what the weather will do to yields, which produce travels well and which doesn’t. The decision tables captured the substitution rules, but Dave’s instinct about what will grow well next month isn’t in any table.
Marcus, the sales lead, makes the commercial point: “Dave is our moat. Hartland Group can buy Freshly’s distribution network. They can’t buy Dave’s reputation. But if Dave’s reputation is only accessible through Dave, it’s not really a moat. It’s a dependency.”
Charlotte
Charlotte’s assessment surprises her.
The team lists what Charlotte holds: the scaling methodology, the cadence design, the quarterly planning framework, the connection between discovery techniques and business outcomes. The way she adapted the planning onion for GreenBox’s specific context. The retro questions. The squad structure principles. The entire rhythm of how twenty-five people coordinate across three cities.
“It’s all documented,” Charlotte says from the phone. “The cadence is on the wall. The planning onion is on A3 paper. The squad structures are in the wiki.”
Diane nods. “The process is documented. The judgement isn’t. When a squad lead comes to you and says ‘this doesn’t feel right,’ you know whether the answer is ‘push through’ or ‘stop and rethink.’ That judgement is worth more than the process.”
Charlotte goes quiet. She’s thinking about the meal kit company – the one she scaled before GreenBox. The processes survived her departure. The judgement didn’t. The company was acquired six months later, and the acquirer threw out every practice Charlotte had built because nobody could explain why they mattered.
“You’re right,” she says. “I need to teach the why, not just the what.”
The mitigation plan
Patricia takes over. “We’ve identified the risk. Now we reduce it. Every key person risk has the same four mitigations.”
Documentation. Get what’s in people’s heads into systems. Not as a one-off project – as an ongoing practice. Maya’s customer relationships go into the CRM with context notes. Tom’s architectural narrative becomes a living document alongside the ADRs. Charlotte’s methodology rationale gets recorded alongside the processes. “You’ve already done this for domain knowledge with the decision tables,” Patricia says. “Now do it for everything else.”
Cross-training. Every critical function needs at least two people who can perform it. Not as well as the primary – just well enough. Sam shadows Maya on farm visits. Priya and Kai learn the full deployment process. Anika starts leading cadence reviews alongside Charlotte. “The backup doesn’t need to be as good. They need to be good enough that the company survives six months while you find a permanent replacement.”
Relationship transfer. This is the hardest one. You can document a process. You can cross-train a skill. You can’t transfer a relationship. But you can broaden it. Maya starts bringing Sam to farm partner meetings. Ben is already the day-to-day contact for Dave’s farm. Marcus begins building his own relationships with key subscribers. The goal isn’t to replace Maya’s relationships – it’s to make sure the company has multiple relationship holders for each critical partner and customer.
Actual insurance. GreenBox takes out key person insurance policies on Maya and Tom. The insurance doesn’t prevent the loss – it provides financial runway. If Maya or Tom were unable to work, the payout would cover twelve months of replacement hiring and transition costs. It’s a safety net, not a solution. But it’s a safety net the company doesn’t currently have.
Patricia summarises: “The goal isn’t to make anyone replaceable. The goal is to make the company resilient. There’s a difference. Replaceable means someone else can do your job. Resilient means the company can survive while it figures out what to do next.”
Dave’s farm
Two days later, Maya drives to Margaret River.
She hasn’t called ahead. She never does. Dave doesn’t answer calls from numbers he doesn’t recognise, and he screens calls from numbers he does. The easiest way to talk to Dave is to show up.
She finds him in the shed behind the main house, servicing the irrigation pump. Ben is in the field, running the new sensor system that Kai helped install six months ago. Helen waves from the kitchen window and brings out tea without being asked.
Maya and Dave sit on upturned crates in the shed – similar to the position they were in at the farmers’ market when Lee visited in the planning onion post. The morning is cold. The kind of Margaret River cold that sits in your bones.
“The board did an exercise,” Maya says. “About what happens if key people disappear.”
Dave puts down his wrench. “Key people meaning you.”
“And you. And Tom. And Charlotte.”
Dave is quiet for a moment. “What did they decide?”
“That we need to reduce the risk. Document things. Cross-train people. And –” she hesitates. “Put some things on paper that have always been handshake deals.”
Dave looks at her. His expression doesn’t change, but something shifts behind his eyes. Maya has seen that look before – when Freshly called him and offered guaranteed volume. When the decision tables codified his substitution knowledge. It’s the look of a man who built his life on trust being told that trust isn’t enough.
“We have a handshake, Maya. That’s how it’s always been.”
“I know. And I value that more than any contract. But the board needs to know that if either of us isn’t here, the relationship survives. It’s not about us. It’s about what happens after us.”
Dave looks out the shed door at the paddock. Ben is walking the rows, checking moisture levels on a tablet. Three years ago, Ben was sceptical of the GreenBox partnership. Now he manages three farm relationships and a supply system that serves twelve thousand subscribers.
“Ben’s the future anyway,” Dave says. “He runs it. I just show up.”
“That’s not true and you know it.”
Dave almost smiles. “No. It’s not true.”
They sit with it for a while. Helen brings more tea. A magpie warbles from the fence line.
“All right,” Dave says. “What does ‘on paper’ look like?”
They work it out over the next hour. Not a corporate contract – Maya knew better than to bring one. A simple partnership agreement, handwritten by Dave on lined paper from the kitchen drawer, typed up later by Sam.
The agreement covers three things: GreenBox commits to purchasing a minimum volume from Morrison Farm each quarter, adjusted seasonally. Morrison Farm commits to GreenBox as a priority customer. And either party can exit with ninety days’ notice.
Dave insists on one clause. He writes it himself, in the careful block letters of a man who doesn’t trust cursive with important things: “If GreenBox ever stops buying local, the agreement ends.”
Maya agrees without hesitation. She would have insisted on it herself.
Dave reads the whole thing back. Then he looks at Ben, who has come in from the field. Ben nods.
Dave signs it. Not because the board told him to. Because Ben is going to inherit this relationship, and Ben deserves something more than a handshake he wasn’t present for.
The knowledge capture sprint
Back in Perth, the mitigation plan becomes a two-week sprint. Charlotte calls it “the knowledge capture sprint” – a period where the team’s primary goal is to get critical knowledge out of people’s heads and into shared systems.
Tom writes an architectural narrative. Not more ADRs – a single document that describes how the system fits together, why the major boundaries exist where they do, and what he would tell a new engineering lead on their first day. It takes him three days. When he’s done, he asks Priya to read it and note everything that’s missing or unclear. She returns it with forty-two annotations. They spend an afternoon resolving them.
Maya records context notes for every farm partner relationship. Not just contact details and order history – the personal context. Dave’s frost story. Rachel’s broadband problems. The new partner in Melbourne whose wife just had a baby and who needs flexible delivery scheduling. The context that makes the relationship human, not transactional.
Charlotte creates “methodology rationale” documents for each of the key practices she introduced. Not process descriptions – those already exist. Rationale: why this practice, in this order, for this type of company. What to watch for. What failure looks like. When to stop doing it. She asks Anika to review them, because Anika is the person most likely to carry the methodology forward.
Sam documents the operational playbooks that exist only in her head. What to do when a Thursday delivery fails. How to handle a customer complaint about produce quality. The escalation path when a farm partner can’t fulfill an order. The seventeen things Sam does every morning before anyone else is in the office.
The knowledge capture sprint surfaces an uncomfortable truth: the team has been operating on institutional knowledge that was never institutional. It was personal. It lived in the heads of five people who were present from day one. Everyone who joined later – Kai, Ravi, Anika, Marcus, the Melbourne and Brisbane teams – has been operating on second-hand understanding, absorbed through osmosis rather than documented and taught.
Tom puts it best, in a message to the team Slack channel at 11pm on a Thursday: “We spent three years building systems to capture customer knowledge – decision tables, JTBD insights, assumption maps. We never built systems to capture our own knowledge. The irony is not lost on me.”
Ren
The following Saturday, Maya meets Ren Tanaka on the Fremantle waterfront. Their usual spot – a bench near the fishing boat harbour, facing west toward Rottnest Island. It’s mid-November. The evening light is doing the thing it does in Fremantle, turning everything gold and long-shadowed.
Ren brings tea in a thermos. She always does. Maya has never seen Ren drink coffee.
They sit for a while without talking. Ren is good at silence. She’s the only person Maya knows who treats silence as a complete sentence.
“The board did an exercise,” Maya says eventually. “Bus factor. What happens if key people disappear.”
Ren pours tea. “And?”
“Apparently I’m the biggest risk to my own company.”
“You’ve been the biggest risk to your own company since day one. The difference is now you know it.”
Maya laughs – the kind of laugh that’s half relief and half exhaustion.
“I went to see Dave. About formalising the partnership. He signed something. First time in three years.”
“How was that?”
“Hard. Good. He did it for Ben, not for the board.”
Ren nods. “That’s the right reason.”
They watch a fishing boat come in through the harbour mouth. The skipper waves at someone on the jetty. The evening crowd is starting to gather along the cappuccino strip.
“I’ve spent the last two weeks documenting everything in my head,” Maya says. “Customer relationships, farm context, strategic thinking. All of it. Putting it in systems so the company doesn’t depend on me.”
“How does that feel?”
Maya thinks about it. “Like selling the farm.”
Ren looks at her. She knows what the farm means. She’s the person Maya called at 2am when her father told her the farm had been subdivided. She’s the one who drove Maya to Margaret River the next morning and sat in the car while Maya walked the fence line of what used to be her parents’ paddock.
“It’s not selling the farm,” Ren says. “It’s planting seeds in more than one paddock.”
Maya stares at the water.
“What are you going to do about it?” Ren asks.
“About what?”
“About being the biggest risk. You know the problem. You’ve started the mitigation. But you haven’t answered the real question.”
Maya waits.
“The real question is: what kind of company do you want? One that needs you, or one that works without you?”
“Can’t it be both?”
Ren pours more tea. “Ask your mum.”
Maya is quiet for a long time.
The management layer is forming. The governance is in place. The key person risks are being addressed. Now Maya needs to stand in front of investors and make the case. The pitch is in three weeks.