All Gold(wo)men Were Once Greenhorns

23:47. Trace Flint, 23, sits alone on the 17th floor. The floor is hollow—empty seats, darkened screens, the faint whir of servers below. Her screen glows: Subtrax_Model_v44.xlsx. Four years of company financials. Line items stacked thirty-deep. The model was built by an intern two months in. The intern was smart. The intern went to Oxford. But the intern made a fundamental error in his cost forecasting.

She scrolls through the detail tabs. There—in "Cost Structure Assumptions." Support costs: marked as "Fixed at £187k." She writes in her notebook: "Support scales with headcount. Headcount scales with customer count. If you double customers, you roughly double support tickets. You need more support staff. Cost is not fixed."

She scrolls down to the summary. The intern's model shows: Revenue growing from £5.3m to £16.6m by FY28. EBITDA turns positive in year two—£4.2m, climbing to £9.8m by year three. Too smooth. Too neat. When you add customers, costs don't stay flat. The cost rises with headcount. You can't keep support costs flat while your customer base triples.

She clicks into the assumptions and rewires the mechanics. Support costs as 5% of revenue instead of fixed. Host costs scaled more realistically. The new trajectory is radically different. Revenue still grows to £16.6m. But EBITDA: negative in year two (now -£1.1m instead of +£4.2m). The inflection point is delayed. By year four: £3.2m, not £9.8m.

She makes a note in clean handwriting: "Numbers fixed. Cost structure corrected. EBITDA trajectory more realistic per actual cost scaling. — TF"

She saves the file as "Subtrax_Model_v45_CORRECTED" and emails it to herself. She closes the lid and walks to the DLR. She thinks about the model. Tomorrow someone will take her corrected version and build a pitchbook. That pitchbook will form a valuation. That valuation will determine whether Subtrax gets funded or not. She doesn't know this yet. But her three-hour correction is now doing real work in the world.

Le Pitch stands on the balcony with a glass of eau-de-vie. The sky above La Défense glows faintly orange. He has just finished a phone call with Haircut—casual, "checking in," but the message was clear: "Interessant?" Yes. Très interessant. He pulls his notebook and writes: "Projet Forge — KM — initiate dual-track. Le Pitch lead. Trace on the ground." He underlines "Trace." Tomorrow he will call her. Tonight, Puccini plays. La Bohème. The moment where Mimi and Rodolfo realize they are falling in love, and the orchestra swells, and you know everything is about to change. That is what the beginning of a deal feels like. Tomorrow belongs to the bank. Tonight belongs to this moment—the moment before everything accelerates.

Trace is packing her bag when Le Pitch's email pings: "Brilliant work on Subtrax. See me 7am before market open. Something new." The 17th floor is completely empty. She takes the DLR home alone. Yellow streetlamps blur past. She thinks about the model she corrected—the one with the wrong cost structure. The intern treated support costs as fixed. As if a doubled customer base required no more infrastructure. Costs don't stay flat. They scale. Growth isn't linear. It's a series of inflection points where old assumptions break. She pulls out her notebook and writes: "The numbers tell the story. The story tells where you go next." She closes the notebook. Tomorrow, Le Pitch will call her into his office. She will hesitate for half a second—thinking of her brother's exams—and then she will say yes.

Haircut sits alone, coffee cold on the desk. On the screen: "Projet Forge — KM Teasers." "KM Financials — Preliminary." "KM Valuation — First Pass." The numbers are clean but the business is not. Asset-heavy, cash-hungry for working capital, cyclical demand, owner in succession crisis. Entry multiple around 8x EBITDA of €10m = €80m enterprise value. If they can grow volume by 10%, EBITDA hits €12m. Exit multiple 10-11x = €120-132m. Equity gets €60-72m after debt. Return of 1-1.2x. Probably 15-18% IRR if volume stays stable. But volume won't stay stable. He opens a blank spreadsheet. "Base case: steady growth, 15-18% IRR." Then: "Downside case: volume -20%, need to refinance." He sits back. He is maybe 40% confident. Sixty percent is hope. Below, he writes: "The numbers don't lie. People do. Wilhelm is proud. Pride can be expensive."

Wilhelm sits at his oak desk—his grandfather's desk. The sealed letter to Matthias waits in the postbox by the front gate. Red wax, fountain pen. An envelope arrives by courier at 23:47 from Skarn Capital, from Kai Eisenberg (the letter says to call him "Haircut," a detail that irritates Wilhelm). An offer to "discuss strategic options." Wilhelm reads the first page and sets it down. He pours a glass of schnapps from the bottle his father left him in 1987, now almost empty. The letter is polite, professional, and asks the question he has been avoiding for five years: What happens after me? He stands and walks to the window. Snow falls on the factory yard. The machines gleam even in darkness. He places his hand on the glass. It's cold. He thinks about his father. Who took the company through the oil crisis of the 70s. Who refused to cut the apprenticeship program. That was the lesson: loyalty is an investment. So Königshof never cuts. Which is why Königshof has always stayed independent. Which is why he keeps €40m in the bank. Tonight, he allows the question to sit. His hand stays on the cold glass. The snow continues to fall.

Magda closes her laptop. The code commits with a soft ping. Three hours on one database query—800ms to 190ms. It's work no customer will notice. But it will matter. She locks up at 01:12 and takes the Overground home. The train is nearly empty. She reaches her flat and opens her laptop one more time. Revenue: £5.3m run rate. Climbing. Churn: 12%. Falling. The trajectory is mathematically clear: profitability in three years if growth continues, two if they tighten costs. Two things trouble her. One: she is tired. Building used to be pure creation. Now it's management. Metrics. Meetings. Four hours today on a budget meeting. Four hours. She wrote maybe fifty lines of code. Two: the inbound interest. Oracle last month. Salesforce the month before. Stripe wants a call next week. They're not desperate. They're patient. Tomorrow, around 10:30 AM, her phone will ring. A banker she doesn't know. The voice will be carefully polite: "Ms. Kowalska? This is James Chen from SBCI. Two companies, both at inflection points. One asset-heavy, one asset-light. Would you be open to a preliminary conversation?" She will say no. But she will ask for his number. And she will save it. And she will listen to his voicemail three times before deleting it.