Utah · 10 min read
Utah Business, Part 4: Operations That Scale From 3 to 30 People
Part 4 of a weekly series: what breaks at 3, 8, 15 and 30 people, five one-page processes, decision rights with limits, three numbers and a weekly meeting.

A snow-removal and landscaping owner in Utah told me his business had "outgrown him." He was right, but not the way he meant it. In three years he had gone from himself and two friends with a truck to fourteen people and nine vehicles. His phone rang at 4:40 a.m. every time it snowed. Crews called him to ask which driveway came next. Customers called him to ask where the crew was. His wife called him to ask when he was coming home.
"I am the system," he said. "That is the problem."
It was. And it is the most common problem I see between 3 and 30 people. The business did not need a bigger owner. It needed a small, boring, written-down way of passing work from one person to the next, so that the owner could go back to sleep.
This is Part 4 of the series Utah business, run like a project. Part 1 covered the start, Part 2 money and support and Part 3 hiring. Now the thing that decides whether all those new people make the business better or only louder: operations. I am an operations and project specialist, not a lawyer or an accountant, and everything below is my own method and my own reading of what I have seen in small teams. The example business is illustrative, with invented numbers.
What breaks, and when
Teams do not fall apart on a schedule, and the numbers below are rules of thumb from my own work, not statistics. Still, the pattern is so regular that it is worth naming. I see four stages.
About 3 people: everything is in the owner's head. This works beautifully. Nobody needs a process, because the owner is the process. Questions get answered over the tailgate. The danger is not a problem; it is a habit. Everything the owner knows is now a habit that will have to be unlearned later.
About 8 people: the owner's head is full. The first signs are small. Two people ask the same question on the same day. A job gets done twice, or not at all. Somebody says "I thought you were doing that." The fix here is not software. It is a shared list of who owns what.
About 15 people: the middle goes missing. Now there are enough people that the owner cannot talk to everyone every day, but not enough managers to do it for them. Information reaches some people and skips others. Good people get frustrated because decisions take two days. Here you need decision rights: who may decide what, up to what limit, without asking.
About 30 people: nobody sees the whole. Different teams have different versions of the same truth. Sales promises one thing, delivery hears another and the invoice says a third. The fix is a small scoreboard, shared across teams, that everyone reads the same way.
The useful part is the order. You do not fix stage 4 problems at stage 2. Writing a 40-page policy manual for a team of eight is how you get a manual nobody reads and a team that resents it. Fix the problem you have, with the smallest thing that works, and fix the next one when it shows up. That is lightweight governance in one sentence.
The five pages
If you only write down five things, write these. One page each, in plain words, with a named owner who is not always the owner of the company.
- Win the work. How does an inquiry become a yes? Who answers, how fast, what do they ask, what do they send? I took one apart in the process teardown of the new inquiry and the one on the quote.
- Deliver the work. What happens between "yes" and "done"? Who plans it, who does it, who checks it, what does "done" mean?
- Bill the work. How does a finished job become an invoice, and who chases it? Cash is the one thing a growing business cannot run without, and late invoices are the quietest way to lose it.
- Hire for the work. The hiring project from Part 3, in one page: role charter, checklist, first week.
- Review the work. Once a month, what did we learn? A lessons-learned page is enough.
Notice what is not on the list: marketing plans, org charts, mission statements. They are fine things. They are not the thing that wakes you at 4:40 a.m.
How to write a one-page process
Do not sit at a desk and imagine the process. Watch it, once, on a real job, and write what actually happened. Then ask four questions: where does the work wait, where is it redone, where does a person have to ask someone else and what happens when the usual person is away? The process improvement guide shows how to map and measure in two weeks. The page itself has five parts:
- Trigger: what starts it.
- Steps: five to eight, in the order they happen, each with a verb.
- Owner: one name per step.
- Done looks like: one sentence a stranger could check.
- Numbers: one or two things you will count.
If it does not fit on one page, the process is two processes. Split it.
Decision rights: the cheapest raise you can give
Back to the snow crews calling at 4:40 a.m. Why did they call? Because nobody had told them they could decide. Every question went to the owner because the owner was the only person who was allowed to answer.
You fix this with a short written list: who may decide what, up to what limit, without asking. A crew lead may reorder the route when a street is blocked. A crew lead may spend up to $150 on a fix without asking. A crew lead must call before promising a customer anything not on the price list. Three lines. The first morning it snows, nobody calls.
A RACI chart is the formal version of this, and for most small teams it is more than you need. The informal version is a list of ten decisions, each with a name and a limit. The limits matter. People will not use a freedom with no edges, because they cannot tell where the cliff is.
The owner's job does not disappear. It moves up one floor: from answering questions to setting the limits and looking at the numbers. That is a promotion, even if it does not feel like one at 4:40 a.m.
Three numbers, one page, one meeting
At 3 people you can see the business with your eyes. At 15 you cannot. What replaces your eyes is a scoreboard small enough to read in a minute. In Project Focus part 3 I wrote about three numbers that matter. For a team that is growing, I give each team three of its own, plus the same three for the whole business:
- One for speed: how long from request to answer, or from yes to done.
- One for quality: how often something has to be redone.
- One for cash: how long from finished to paid.
Then a weekly meeting of 30 minutes, which I laid out in the weekly rhythm: look at the numbers, pick one problem, leave with a name and a date. The meeting is the engine. Without it, the scoreboard is wallpaper.
Resist adding numbers. A scoreboard with twenty numbers is a spreadsheet, and a spreadsheet is something you open when something is already on fire. For more on this trap, read about dashboards without vanity metrics.
The "vacation test"
Here is the simplest test of whether your operations scale, and I use it with every owner. Imagine you leave for two weeks with your phone off. Walk through the first three days, hour by hour. Who decides what? Who answers the customer who calls angry? Who makes sure the invoice goes out? Where do the keys live?
Every place where the answer is "nobody" or "they would call me" is a gap in a one-page process or a missing decision right. Write the list. It is your to-do list for the next 30 days, and it is better than any consultant's report, including mine.
Do not take the vacation first and run the test afterward. The test is cheaper and it hurts less.
An illustrative example: the fourteen-person snow and landscape company
Everything in this section is invented to show the method. The numbers are not from a real client.
Take our snow-removal and landscaping owner. Fourteen people, nine vehicles, one phone. We spent two weeks watching, not changing. This is what we wrote down.
- A new customer request took 4 days to get a quote, because it went to the owner, who was on a roof.
- Crews called the owner about 12 times on a snow morning, almost always about routes.
- About 1 invoice in 5 went out late, because the paperwork lived in a truck.
- Payment took an average of 38 days after the job.
We did not buy software. We wrote five pages and a list of ten decisions. Crew leads got a spending limit of $150, a route-change right and a short list of things to never promise. Quotes went to one person with a price list, with a target of one business day. A photo of the signed job sheet went to a shared folder the same day. Invoices were sent every Friday by one named person. A 30-minute meeting on Monday looked at three numbers.
Six weeks later the same owner measured again:
- Quote turnaround: 1 day.
- Calls on a snow morning: 3.
- Late invoices: 1 in 20.
- Days to payment: 26.
The business did not get smarter. It got legible: everybody could see what was supposed to happen, and the owner could see what did.
| Measure (illustrative) | Before | After six weeks |
|---|---|---|
| Time to a quote | 4 days | 1 day |
| Calls to the owner on a snow morning | 12 | 3 |
| Late invoices | 1 in 5 | 1 in 20 |
| Days to payment | 38 | 26 |

What to automate, and when
Owners ask me when to buy software. My answer is boring: after the page, not before it. If the process is unclear on paper, software makes it unclear faster and sends you a monthly bill for the privilege.
Once the five pages exist, ask of each step: is it repeated, is it rule-based and does it wait for a person to remember? Those are candidates. Reminders, status updates, recurring invoices, onboarding checklists. The automation examples article has a list, the return on automation guide shows how to check the payback and no-code or custom helps you choose. If you want AI in the loop, the AI series covers where it helps and where a person must stay in charge.
Scaling is a project, with a plan
Treat the move from 8 to 15 or from 15 to 30 as a project. It has a goal, an owner, a schedule and a scoreboard like any other. A simple version looks like this:

- Weeks 1 and 2: watch. Pick one process and follow a real job. Write what happened.
- Weeks 3 and 4: write. One page per process, a list of ten decisions with limits.
- Weeks 5 and 6: run. Use the pages. Fix what does not work. Start the weekly meeting.
- Weeks 7 and 8: measure. Compare the three numbers with the first two weeks. Keep what works.
Eight weeks. Small steps. If you need help seeing the whole, a 90-day sprint is a good frame, and capacity planning shows where the next constraint will be. In Utah, your free SBDC adviser can be a sounding board for this plan. See Part 2 for how to reach them.
Mistakes to avoid
- The binder. Writing a 60-page manual before you know what breaks. Write five pages.
- The hero. Rewarding the one person who always saves the day. If a person must save the day every week, the process is broken.
- Rules without limits. Telling people to "use their judgment" without telling them where the edge is.
- Software first. Buying a tool to fix a process nobody has described.
- Too many numbers. A scoreboard that needs a meeting to explain it.
- No owner. A process with the whole team as its owner has nobody as its owner.
- Waiting for pain. The best time to write the page is when things are calm. The second best is at 4:40 a.m., and it will be a worse page.
Check where you stand
Use the tool below. Tick what is already true in your business today.
What to do this week
- Write the five process names on one sheet of paper and put a name next to each.
- Pick the process that wakes you up most often and watch one real job from start to finish.
- Write that page: trigger, five to eight steps, owner, what "done" looks like and one number.
- List ten decisions you answer every week and give each one an owner and a limit.
- Choose three numbers (speed, quality, cash) and start a 30-minute weekly meeting.
- Run the vacation test on paper and write down every gap.
Next in the series, Part 5 looks at growing with the region: the 2034 Games, construction and tourism. For a first look at the region's plans, read what the 2034 Games mean for business.
This post is general information about operations and project management, not legal, tax or employment advice. The example business and all its numbers are illustrative. I did not use any customer data. If you want help turning your operations into five pages that run without you, see my services or take the diagnostic.
Frequently asked questions
At what size does a small business need written processes?
There is no magic number, but a common signal is when the same question is asked twice in a day, a job is done twice or missed, or the owner cannot take a week off. In my experience this tends to appear around 8 people, but it can happen sooner. Start with one page for the process that causes the most trouble.
How many processes should I document first?
Five: winning the work, delivering it, billing it, hiring for it and reviewing it. One page each, with a named owner. Add more only when a new problem shows up that those pages do not cover.
How do I delegate decisions without losing control?
Write a short list of the decisions you answer most often, and for each one name who may decide and the limit, for example a spending amount or a type of promise. Stay in charge of the limits and the numbers, not every answer. Review the list every quarter.
Should I buy software before I scale?
Usually not first. Describe the process on one page, run it by hand for a few weeks, fix it and then ask which repeated, rule-based steps could be automated. Software on top of an unclear process makes the confusion faster.
Everything new, once a month.
One email on the 1st of every month with everything new: articles, series parts, news, website updates, project results, where I am working and the business opportunities I am looking at. No spam.
- Fresh on the site right now
- ArticleMeasuring the Return of Automation: Hours, Errors and Cash, Counted Honestly
- NewsSix North Temple murals, $150,000: what a project manager sees in the city's mural program
- SeriesOperations that scale from 3 to 30 people
Join the conversation
Did this help? A quick reaction tells me what to write next.
Be the first to react.

Loading comments…