The job you're proudest of may be the one losing you money

Quick answer: Write down the hours every job takes, including the driving, the quoting and the callbacks. Divide what the job earned, after materials, by those hours. Do it for two weeks and you'll know which work to chase and which to price up or turn down.

The kitchen that felt like a win

Take a made-up but familiar example. You run a three-man remodeling outfit. You quote a kitchen refresh at $9,000. Materials come to $3,200, so $5,800 is left to pay for labor and keep the lights on. You figured 120 hours, which works out to about $48 an hour. Fine.

The same month you do a small bathroom for $4,200. Materials are $1,100, leaving $3,100. You figured it for a quick filler job.

The kitchen was the one you bragged about. It was big, the customer was friendly, and the truck sat outside for three weeks. But when you add it up, the kitchen took 190 hours. Fourteen of them were runs to the supply house. Eleven were callbacks for a cabinet door and a leaky fitting. The customer also asked for "one more thing" six times, and you did all six for free.

$5,800 over 190 hours is about $30.50 an hour.

The bathroom took 52 hours. $3,100 over 52 is almost $60 an hour.

The job you were proud of paid half what the filler job did. Without numbers you'd have gone looking for more kitchens.

Why your gut can't do this

Busy feels like profitable. A big job gives you a full calendar, a good story and a customer who tells their neighbors. None of that shows up in the bank until the month closes, and by then the work is done and the price is set.

Memory is also biased. You remember the day the tile went down perfectly. You forget the three evenings spent on the phone sorting out a delivery.

Tracking time fixes this by taking opinion out of it. It's the most basic form of knowing your own business. If you already look at your numbers weekly, as in The weekly numbers that keep a small business out of trouble, this adds the one number most owners are missing, which is what an hour actually earns.

How to set it up in a day

Keep it dull. If it takes more than a minute a day per person, it won't last past the first busy week.

• Track by job, not by task. You don't need to know how long the grout took. You need to know how long the whole bathroom took.
• Make buckets for the unbillable stuff. Quoting, driving, supply runs, callbacks, admin. Those hours are real and somebody's paying for them. It's you.
• Pick one tool and stick to it. A pocket notebook works. A free timer app on your phone works. A shared sheet that everyone fills in at the end of the day works. A fancy system nobody opens doesn't.
• Set one daily cutoff. Everyone logs their hours before they leave the site or before dinner. After a day, nobody remembers. After a week, they're guessing.

Run it for two weeks, or for your next five jobs if jobs are longer than that.

Do the math on every job

For each job, take the price, subtract materials and any subcontractors, and divide by total hours. Include every hour you and your crew put in, not just the ones on site.

Then compare the result to one number: what an hour of your time has to earn for the business to survive. That's your payroll, your truck, insurance, tools, software and your own paycheck, divided by the hours you can realistically sell in a month. If you're not sure how to build that number, sit down with your bookkeeper or accountant once. It's an hour well spent, and reading your profit and loss in ten minutes a month gets you most of the inputs.

Now sort your jobs into three piles. Above the line, near the line, below the line. You'll probably see a pattern within a week.

What you'll usually find

Patterns repeat across trades and shops. The cause is almost never that the crew is slow.

Your estimates are light on the same thing every time. It could be demolition, cleanup, permits or drive time. Once you see that the third job in a row ran 20 hours over on the same phase, fix the estimate and move on.

One customer eats the schedule. They call daily, change their mind and want a walk-through every morning. Their job might look fine on the invoice. In hours, it's a disaster. Get changes in writing and price them. You can be kind about it and still hold the line. "Happy to add that. It's $340 and two extra days." That's service, not rudeness.

Small jobs beat big ones. This surprises people. Small jobs have less scope creep, less waiting and fewer surprises.

Callbacks are a hidden tax. Eleven hours of free repair work on a $5,800 job is nearly 10 percent of the margin gone. Track them in their own bucket. If one crew member generates most of them, that's training, not blame.

Speed has a cost. Sometimes the fastest crew member has the most callbacks. The numbers show it.

Hold the line without losing the customer

When a job runs under water, you have four options. Reprice future jobs of the same type. Change what's included. Fix the process that's costing you hours. Or stop taking that kind of work.

You don't need to do all four. You need to pick one on purpose. If you want help raising prices without sounding like a salesman, read Selling without the sales voice.

Be fair about one thing. Some low-rate work is worth doing. An anchor customer who sends you six referrals a year may be worth a thin kitchen. A slow February job keeps the crew paid and together. That's a decision. The danger is the thin job you didn't know was thin.

Don't use it as a whip

Your crew will see this as spying if you let them.

Tell them plainly what it's for: it measures jobs, not men. Nobody's paycheck depends on one day's log. Show them the results too. When the numbers prove that the kitchen needed a better quote, and the crew was fine, they'll trust the system. When they see the bathroom paid $60 an hour, they'll understand why you're picking jobs more carefully.

If it's ever used to punish an honest number, they'll start writing dishonest ones, and then you've paid for a tool that lies to you.

Look after your people first. Pay them fairly, give them clean work, and don't blame them for a bad estimate that was yours.

The first time you say no

Around the third week, a quote request will come in for another big, flattering, messy job. You'll pull up the numbers on jobs like it. You'll know what it really pays.

That's when you either raise the price by 25 percent or tell them you're booked. Do it once, and you'll never go back to guessing.