Audit. Build. Run. — What Actually Happens.

The three-word version of the method is on the homepage. The thirty-page version lives inside our client folder. This is the middle version — long enough to be honest about what happens, short enough that you can read it in a sitting and decide if you want the deeper conversation.

To make it concrete, walk through it with a composite client. Call him D. He is 41, owns a services business doing somewhere in the high six figures, works most days, has operated for nine years, restructured the company two years ago, has been chasing the same margin number for fourteen months. He works long hours, handles the hardest accounts personally, travels for one event a quarter, and has a four-year-old. He emailed us in March. Below is how each phase ran for him. Names and details are altered — the shape is real.

Why we don't start with a plan

The first instinct most consultants have when a founder shows up is to write a strategy. That is the part of the job that looks like the job. It is also where most engagements quietly fail, because the strategy is being written against a guess about the bottleneck rather than a measured one.

When D got on the discovery call, his self-diagnosis was that his pricing had gone stale and he needed to raise rates. He had been running a high-volume, low-margin book for the previous year. He wanted a repricing plan. A consultant without an Audit phase would have written it, charged him, and watched the margin stay flat twelve weeks later for the same underlying reason.

We refuse to write a plan in the first week. The cost of refusing is friction with people who want action. The benefit is that we stop solving the wrong problem.

What the Audit actually surfaces

Day 1 through Day 10 of D's engagement was the Audit. He sent us nine months of order, delivery, and invoice records from his tool of choice, exported as CSV. He filled out an 80-question intake covering operating history, current workflows, team structure, tooling, prior advisors, prior reorganizations, and the gap between his stated goals and his last 12 weeks of behavior. He gave us read access to his PM and billing tools for the previous 90 days. We pulled cycle time, exception rate, queue depth at each stage, and cost per unit.

We held a 60-minute Audit call on Day 4. Most of the call was him talking and us writing. Twenty minutes in, the actual bottleneck surfaced in passing. He mentioned, without thinking it was important, that he had not reviewed his project margins in nine months because he kept feeling like he was about to turn the corner. He also mentioned that his most senior person handled the hardest accounts entirely off-process, and that he had stopped running the weekly ops review about a year ago because he could not remember why he started doing it.

Day 8, we delivered the Audit document. Eleven pages. Three prioritized gaps. The first was process debt — nine months of accumulated workarounds driving the margin erosion and, more importantly, driving the perceived need to reprice. The second was a measurement gap — he was capturing inputs but not aggregating them weekly, so the margin leak had been invisible to him for months. The third was an unowned handoff — the senior person's off-process work was producing the company's worst-margin outcomes and nobody was measuring it.

His own diagnosis of "need to reprice" was nowhere in the top three. That was the whole point.

What "Build" means — and what gets cut

Day 11 through Day 18 was the Build. The product was a 22-page operating document. It contained four things.

A 12-week workflow build, structured as a documented delivery process with a forced review at week four. Each stage is gated by an explicit "done" check. The senior person's off-process accounts are moved onto the standard path for the duration of the build. Total work-in-progress is capped about fifteen percent below his previous load, which felt wrong to him until we walked through the margin numbers.

An ownership and handoff map with a defined owner per stage, a handoff contract, an entry/exit check, and a rule for what happens when a handoff drops. No new tool. No new platform. The data was already in his existing stack; he just was not reading it.

A measurement loop — daily inputs, weekly aggregates, monthly recalibration, quarterly review. A Google Sheet, not a proprietary platform. (See the Measurement Stack article for the full structure.)

An accountability cadence: weekly written check-in by Friday 5 PM, response from us by Tuesday 9 AM, monthly call on the first Tuesday, quarterly review document at week 12.

Things that did not make it into the Build: a repricing model, a headcount plan, a new-tool migration, an aggressive cost-cut, or anything in the fundraising-or-financing category. He asked about all five at various points. The Audit had not flagged any of them as the leverage. We declined to write them. This is the part of the practice that frustrates new clients and saves the engagement.

The Build kicked off with a 75-minute onboarding call. We walked through every page of the document, set up the measurement sheet, ran his first weekly review live, and answered questions. Day 18, he started Week 1.

How "Run" prevents the slow drift that kills most engagements

Run is the unglamorous ninety percent. It is also where most independent operators and most consultant-client relationships quietly fail, because the failure is invisible week to week and obvious only at the three-month mark.

D's first six weeks looked like this. Weekly check-ins on Friday — a structured form covering work completed, cycle-time distribution, exception rate, queue depth, process adherence, and one open field for anything off-pattern. Tuesday morning he had a written response from us: what we observed in his numbers, what one adjustment was being made for the upcoming week (tighten the entry check on stage two / hold work-in-progress / move the senior person's next account onto the standard path), and a short note acknowledging whatever was off-pattern.

In week three he asked to skip the forced review. We declined. He held the review. In week six he hit a new margin high on a delivered project, the first in fourteen months. He almost did not notice because he was so focused on the long-term arc.

Week eight, we had the first monthly recalibration call. Thirty-eight minutes. We looked at the four-week arc: planned versus actual work, throughput progression, exception trend, behavioral adherence. We made two small changes — increased the cadence of the stage-two review from weekly to twice-weekly, and added a Wednesday backlog sweep. We did not make five changes. Discipline of one change is the rule.

Week twelve, we ran his quarterly review document. Eight pages. It compared his current data against the Day 1 Audit. The three priority gaps had each closed measurably. Process debt was paid down; exception rate was down eleven percent from baseline. The measurement loop was running on its own; he had not missed a Friday review in twelve weeks. The senior person's accounts, now on the standard path, were the company's most profitable. His project margin was up materially.

What he wrote in the review intake field was the line we hear most often from clients who stay: "I cannot believe I tried to fix this by repricing for fourteen months."

Where this falls apart (and when to walk away)

The method does not work in every case. It is worth being honest about where.

It does not work if the client cannot commit a real thirty minutes on Friday for the review. The cadence is the system. A client who is too busy for the Friday review is too busy for the engagement. We end engagements over this and the people we end them for are usually relieved.

It does not work if the client is fundamentally not ready to be advised — wants validation rather than feedback, treats the weekly check-in as a confessional, or argues with the data when the data is inconvenient. We catch most of this on the discovery call and decline the engagement before any money changes hands.

It does not work for pre-revenue founders. If you have not yet found a product people pay for, you do not need a custom operating system. You need a product, time, and a willingness to be patient. We will tell you this on the discovery call and recommend better-fit resources instead. This costs us business and it is the right answer.

And it does not work if the Audit reveals that the actual bottleneck is something we cannot help with — a legal exposure, a tax problem, a financing gap, or a co-founder situation that needs different professional support first. When that happens, the engagement does not start. If Audit reveals you don't need us, we refund and recommend an alternative. That is the 14-day window on the refund policy and we honor it.

That is the practice. It is mostly boring. The boring part is the point.