How to start making your business run without you

You know the driveway call. A technician, a biller, or a coordinator is standing in front of a live customer problem, and instead of handling it they call you, because you are the only person who knows what the company would do. Every one of those calls tells you the same thing: the knowledge that runs your business still lives in you, and the business cannot grow past your capacity to answer the phone.

The fix has a name: Michael Gerber called it building your business like a franchise prototype, and our E-Myth Notes series covers the idea in depth. This post is not about the idea. It is about the first project: what to actually do, starting this week, with the business you already have.

One warning before the steps. This project is designed against the binder trap: spending three months documenting every process in the company from memory, and producing a shelf of SOPs nobody opens because they describe the work as the owner imagines it rather than as it happens. Do not start with everything. Start with one workflow and take it all the way to a successful handoff. The handoff is the unit of progress, and the proof.

Step 1: Pick the workflow that interrupts you most

Not the most important workflow. The one that generates the most driveway calls: recurring, customer-facing, and currently dependent on your judgment. For an HVAC company it might be the no-cool service call. For a dental office, insurance verification before treatment. For an agency, the scope-change conversation.

You are looking for a workflow that runs at least weekly (so you get enough cases to learn from), follows a recognizable shape most of the time, and hurts when it goes wrong. Write down its name and the promise it makes to the customer. If you cannot state the promise in one sentence (“same-day diagnosis and a price before we leave”), that is your first discovery: the workflow has no defined finish line, which is why every case comes back to you.

Step 2: Capture the next five real cases

For the next week or two, every time this workflow runs, write down what actually happened, during or immediately after, never from memory at the end of the week. Memory smooths out exactly the details you need. For each case, capture: what the customer asked for, what information the person handling it had at the start, every decision that got made and by whom, every moment someone had to ask you something, and how the case ended.

The moments someone had to ask you are the gold. Each one is a rule that exists only in your head. Five cases will surface most of the common ones.

Step 3: Write the one-page runbook

Now — and only now — write the runbook, from the cases rather than from your self-image of how the work goes. One page. If it does not fit on one page, the workflow you picked is really several workflows; split it.

Here is a filled-in example for the HVAC no-cool call, so you can see the altitude to aim for:

Workflow: Residential no-cool service call Promise: Diagnosis today; a written price before any repair; no repair over $600 without the customer hearing options.

Before dispatch: confirm the address is inside our service area, confirm the unit’s age and last service date from the customer record, tell the customer the diagnostic fee on the phone.

On site: run the diagnostic checklist (page 2). Photograph the data plate and any failed part before touching it.

Decision limits:

• Repair under $600 with parts on the truck: proceed on customer OK, no call-in needed.
• Repair $600–$1,500: present the written option sheet; customer signs before work starts.
• Repair over $1,500, or unit older than 12 years: present repair vs. replace using the option sheet; schedule the comfort advisor and do not quote replacement pricing on site.
• Anything not on the option sheet: call the service manager, not the owner.

Every case ends with: photos attached, option sheet (signed or declined) attached, next step recorded (done / parts ordered / advisor scheduled), and the customer told what happens next before we leave the driveway.

This runbook fails when: the customer disputes a price we never wrote down, or a case ends with no recorded next step. Both go to the Friday review.

Notice what the page contains: the promise, the facts to gather before acting, numeric decision limits, an escalation path that is not you, and a definition of what “finished” looks like: evidence attached, next step named. Notice also what it does not contain: technical instructions for diagnosing a compressor. Your technician knows the craft. The runbook captures the business judgment around the craft, because that is the part that currently lives only in you.

Step 4: Hand it off, and only intervene through the record

Give the workflow to the person who should own it, with the runbook, and let the next five cases run. Two rules for you, and they are harder than they sound.

First, when they call you anyway, answer, and then add the answer to the runbook the same day. Every call is a missing rule, and the runbook is where it goes to die as a future interruption.

Second, review cases through the record, at a set time, instead of hovering. A short Friday review of the week’s cases against the runbook: which decision limits got used, which case ended without a next step, what the customer disputed. If you find yourself correcting work in real time all week, the runbook is not done, or the decision limits are set at the wrong thresholds. Fix the page, and let it run again.

How you know it worked

The test is blunt: did someone else run five consecutive cases, to the promise, without a driveway call? When yes, you have converted one piece of private judgment into company property. Pick the next workflow; you now also have a team that has seen the pattern work, which makes the second one faster.

Be honest about what this project does not do. It will not fix a broken pricing strategy, hire your service manager, or resolve whether you are in the right market. Our judgment (and it is judgment, drawn from the E-Myth tradition and our own operating work rather than from a controlled study) is that the sequence above is still the highest-leverage first move for an owner who is the bottleneck, because every later fix gets easier once work leaves your head and starts leaving evidence.

A business becomes sellable, delegable, and calmer the same way: one workflow at a time, proven by handoff. Start with the one that will ring your phone tonight.