# Unstable by design _You can choose instability for its advantages, but your ability to respond must be faster than the system's tendency to diverge._ --- In 1917, the Sopwith Camel entered service with a problem its designers had put there on purpose. The engine, fuel tank, guns, and pilot were all packed into the front seven feet of the airframe, a concentration of mass that made the aircraft inherently nose-heavy and rotationally unstable. Left to itself, a Camel didn't fly straight. It wanted to pitch and yaw with every gust, every throttle change, every shift in the pilot's weight. This was the point. A stable aircraft resists being turned. An unstable one is already halfway into the turn before the pilot moves the stick. In a dogfight over the Western Front, where a fraction of a second determined who got behind whom, the Camel's instability was a weapon. It could flick into a right-hand turn faster than anything else in the sky. But the same physics that gave it agility also gave it a kill rate among its own pilots that rivalled enemy fire. The Camel demanded constant correction just to hold a heading. If a trainee pilot overcorrected on a turn, the torque from the rotary engine would tighten the spiral, and the aircraft would spin into the ground before most hands were fast enough to recover it. More Camel pilots died in training than in combat. The instability that made it lethal in a dogfight made it lethal in every other circumstance too. --- You've just joined a company that runs a leveraged roll-up strategy. The thesis is clean: acquire small, profitable businesses in a fragmented market, fund them partly with debt, apply operational improvements, use the growing cash flow to service the borrowing and fund the next acquisition. Each deal is accretive. The maths compounds. On paper, the model looks like a flywheel. Three months in, you can see why it works. The leverage amplifies returns on equity. You're buying businesses at five or six times earnings using capital that costs you four percent, and the spread drops straight to shareholder value. The acquisitions bring revenue that supports more borrowing, which funds more acquisitions. Competitors running the same strategy without leverage grow half as fast. You are, in the language of the Camel's designers, deliberately unstable, and the instability is giving you speed. --- Then a deal slips. One of the recently acquired businesses misses its first quarter under your ownership. Revenue is down eight percent, which is modest in isolation, but the covenants on your credit facility are written against consolidated EBITDA. The miss pulls your leverage ratio from 3.1x to 3.4x, close enough to the 3.5x ceiling that your lender calls to ask questions. The questions aren't hostile, but they change the atmosphere. Your CFO spends a week preparing covenant compliance materials instead of working on the integration plan for the next acquisition. The next deal, which was in final due diligence, gets paused because the board wants to see how the ratio settles. Your pipeline, which depended on closing two deals a quarter to hit the annual plan, is now one deal behind. In a stable business, a single quarter's revenue miss is a line item in the board pack. Unpleasant, manageable, absorbed. In your business, the miss doesn't stay where it lands. It travels. Revenue shortfall tightens covenants, tighter covenants slow the acquisition pipeline, a slower pipeline means less revenue growth to service the existing debt, and less revenue growth pushes the ratio higher still. The error compounds in a direction that makes the next error more likely. --- This is the same physics as the Camel, translated into a balance sheet. The leverage that amplifies your returns also amplifies your deviations. Small wobbles in a stable system damp themselves out. Small wobbles in your system feed back into themselves and grow. You could see this coming if you tracked the right number. Every leveraged model has a rate at which problems compound, a speed of divergence set by the interaction between your debt load, your covenant headroom, and your revenue volatility. Your response loop, the time it takes to see a problem, decide what to do, and execute, has to be faster than that rate. If it is, you can correct before the wobble becomes a spiral. If it isn't, you're the trainee pilot watching the ground rotate towards you. --- So you look at your response loop. Monthly management accounts take three weeks to close. That means you're flying on instruments that show you where you were a month ago, not where you are. The acquired businesses report on different cycles, some weekly, some monthly, with varying definitions of what counts as revenue. Consolidation takes another week. By the time you see the miss, it's already six weeks old. Decisions move at a similar pace. The acquisition committee meets every two weeks. Any deal above a certain size needs board approval, which means scheduling around non-executive diaries. Covenant conversations with the lender require the CFO, the CEO, and external counsel, none of whom are easy to assemble at short notice. The fastest you can move from "we see a problem" to "we've done something about it" is roughly eight weeks. Eight weeks is fine for a company running at 1x leverage with stable revenues and no covenants to trip. It is catastrophically slow for a company running at 3x leverage where a single quarter's miss cascades through the capital structure. The instability of your model demands weekly visibility, pre-agreed responses to covenant pressure, and [[Reversible decisions]] that don't require assembling a committee. You have none of these. --- The appeal of the leveraged roll-up, and the appeal of the Camel, is the same: instability as advantage. You take on risk that a more conservative competitor won't, and the instability gives you speed, agility, returns they can't match. That part of the thesis is real. What gets left out of the pitch deck is the control system required to survive it. Your competitors running unlevered roll-ups are slower, but they can absorb a bad quarter without triggering a chain reaction. They fly stable aircraft. Their errors self-correct, or at least don't self-amplify. They have time to think. You don't, and you built the model knowing you wouldn't. The question is whether you also built the instruments, the decision rights, and the reflexes to match. [[Paying for growth]] captures one version of this problem: spending faster than the economics can sustain. Instability captures a broader one. Any system where errors compound rather than dissipate requires a control loop running faster than the compounding. Leverage is one example. Aggressive growth targets that create self-reinforcing expectations are another. Tightly coupled operations where one failure cascades into the next are a third. The mechanism is the same in every case: the gap between your system's speed of divergence and your organisation's speed of response is the gap that kills you. --- The Camel was retired in 1920, replaced by aircraft that were both manoeuvrable and forgiving. The pilots who survived it, though, had learned something that their peers in stable aircraft never had to: how to sense a divergence in its first moment and correct before it built. They flew with a quality of attention that more stable systems never demanded. The company that chooses instability and builds the control system to match it earns something similar. The discipline of fast sensing, fast decisions, and fast correction doesn't just keep you alive. It becomes, over time, a capability your stable competitors can't replicate, because they never had to develop it. But the order matters. The capability comes first. The instability is only a weapon in hands fast enough to hold it.