How the work actually gets done — the steps I run, and the judgements that govern them. The examples are representative, not a fixed checklist.
The same method runs through two kinds of business. A larger business is held together by its systems; a founder-led business is held together by the founder — a natural result of how it grew, not a flaw. What changes between them isn't the method but the work itself: inside a larger business, fixing an engine that already runs; in a founder-led one, often building the engine and freeing the business from the person holding it together. The sequence, and the judgement behind it, stay the same. I work in both.
The Method
The first explanation for a struggling business is almost always about behavior. It's rarely the real cause.
A Commercial Operation ExampleIn Redesigning Value Capture to Plug Margin Leakage — where margin was leaking and every proposed fix aimed at behavior, from discount discipline to sales retraining — each held a symptom, none held the cause: "This was not a behavior problem. The value capture architecture had been left in place through a business model transition that had moved past it."
A Founder ExampleIn a solo coaching practice, flat lead-flow looked like a marketing problem — but "that is not a volume failure at the top of the funnel; it is a value-proposition failure in the middle of it — more leads would only have produced more conversations that went nowhere." The diagnosis forced a pivot from career coaching to EI coaching.
When doing more of what worked before only makes things worse, the problem isn't effort — the business has moved to a new stage.
A Commercial Operation ExampleIn Leading a Category Through Every Transition — where tightening the rules to control pricing produced cleaner failure rather than better results — the tell was that more effort deepened the problem: "More enforcement on pricing produced cleaner failure, not better outcomes. Tighter rules, same variance — the inversion was the signal."
Founder-stage businesses are usually still finding what works, so the inversion comes later.
Test the bet before you back it. You can diagnose a business that already runs — but an untested idea has to earn proof first.
A Commercial Operation ExampleIn Leading a Category Through Every Transition — entering a new category where the obvious move was to scale everywhere at once — the contrarian bet carried its own proof: "Beachheads chosen for visibility, not size — education and mobile workforce first, because deployment at scale in those segments gave later buyers peer evidence instead of a leap of faith," on the principle that "Conviction without risk control is a bet, not a thesis."
A Founder ExampleA clinician-founder — who understood the problem better than almost anyone, and mistook that for having a business — proved people would pay before building anything: "Having identified a real problem is not the same as having a business — the discipline was to prove people would pay before building anything, and to test it outside the people who wanted to be supportive."
Get the design right first. No amount of effort fixes a structure that was never built to work.
A Commercial Operation ExampleIn Organized for Activity, Not Return — a sales force busy and full of pipeline, yet with contribution per rep falling — the lasting change was structural, not motivational: "The durable asset of an organizational redesign is the paradigm, not the structure."
A Founder ExampleIn a pre-launch delivery business, the process and hand-off architecture was built and tested before execution began, so the model could run without improvisation: "A combined system is impossible to diagnose; the discipline was to build and test the architecture in sequence before running it live — operations first, then demand, then pricing — so any failure had a clear owner and a clear fix."
Running two business models through one system forces a deal-by-deal compromise that serves neither. Separate them.
A Commercial Operation ExampleIn Rebuilding a Model That No Longer Matched the Buyer, the split was made explicit — the model "differentiated between strategic enterprise accounts requiring lifecycle-oriented engagement models and transactional accounts optimized around scale, velocity, and acquisition efficiency." The same discipline runs through Leading a Category Through Every Transition, where the resolution was structural separation, not compromise: two engines, two mandates, two sets of metrics, made visible at leadership level.
A Founder ExampleIn a compliance-testing business, walk-in consumer revenue and the intended corporate engine were counted as one number that told a comforting but false story: "The consumer walk-in revenue was masking the complete absence of business-to-business progress; the discipline was to separate the two revenue lines and track them apart, because blending them hid the one number that actually measured the model working."
A business is only really finished when it can run without the person who built it.
A Commercial Operation ExampleIn Organized for Activity, Not Return, the test of the rebuild was whether it held up after the person who built it left — what the successor inherited: "The successor inherited an operating substrate that could absorb the next strategic load, not a turnaround that needed defending."
A Founder ExampleIn Building the Business Around the Opening, value that had built up in the founder was deliberately moved out — deployment, qualification, onboarding and support "all written down as a playbook, so it no longer depended on the founder's memory," until "the real test was whether it could run without the founder. It could."
The Principles Beneath It
Not steps. Standing judgements that govern how the steps get run.
In a closed market, credibility opens the door before anything gets sold.
In Building the Business Around the Opening: "In a closed market, an outsider's first job isn't selling — it's finding a credible way in. Borrow it if you can't build it yet" — because "for an outsider, borrowed credibility was the only credibility available."
The party in control only moves when the new arrangement clearly benefits them. Where it never will, stop trying to convince them and design around them.
A Commercial Operation ExampleIn Turning Software Into the Platform Anchor: "The only way to end that was to make platform standardization more profitable for the GSI than staying swappable. And to build the improvement such that vendor economics improved at the same time, not at vendor cost." The same move drives Creating the Buying Motion in an Emerging AI Category.
A Founder ExampleWith a party controlling the outcome who had no reason to cooperate: "When the party who controls the outcome gains nothing from the change, design around their economics rather than waiting for cooperation their incentives will never give."
Revenue is the number everyone watches. Margin is the one that quietly breaks a transition — track it separately.
In Redesigning Value Capture to Plug Margin Leakage, margin fell through the model transition and recovered only because the redesign was held in place through the early deals: "the redesign stayed in place until the third deal… proved what it was built to prove" — margin finishing 50 basis points above the earlier baseline.
Forecast and rev-ops credibility is a precondition for structural change, not a parallel workstream.
In Organized for Activity, Not Return: "Cross-functional governance built during the credibility phase is the instrument the structural phase runs on."
The person who can kill a deal is often never in the room. Find them — it isn't always who holds the budget — before you back your champions.
A Commercial Operation ExampleIn When the Direct Model Stops Paying for Itself, the redesign moved only with veto power addressed directly: "The architecture would not have moved without senior backing applied repeatedly, by name."
A Founder ExampleA public-sector deal died to veto players who were never in the room: "In institutional sales the deal is decided by who can say no, not who says yes — the veto player must be mapped and neutralized before the champions are ever activated, because the party with the most to lose from the change outranks the party with the most to gain."
Two settings, one method. The work looks different in each because they're built differently — but the sequence, and the judgement behind it, don't change. I work in both.