You're Taking Too Many Jobs — And Losing Money on All of Them
You're Taking Too Many Jobs — And Losing Money on All of Them
By spring of 2008, Smith Mechanical of Worcester was running six trucks and twelve guys. I thought that meant things were good. I was measuring success in trucks on the road — how many rigs in the parking lot Monday morning, how many names on the schedule board. What I was not measuring was whether each truck was actually producing margin. That cost me four men by the end of that year.
Every shop has a throughput ceiling. It's a real number. Most owners have never calculated it. They run on feel — "we're slammed," "we're slow," "I think we can squeeze one more in." That gut read doesn't pay truck notes.
Your Throughput Ceiling Is a Number, Not a Feeling
When I opened out of my garage in 2005, two trucks, I could tell you exactly what each one needed to bill per week to cover its note, fuel, wages, my draw. That number was the ceiling. Everything above it was margin. Everything below it was a problem I needed to fix before Friday.
By 2007, four trucks, I stopped doing that calculation cleanly. There was work, the schedule was full, I was busy — which I confused with profitable. Six trucks and twelve guys, I'd completely lost track of what each truck needed to produce. I knew the total. I didn't know if any individual truck was carrying its weight or being carried.
The second problem was the foreman ladder. I had one working foreman worth a damn — a guy who'd run three guys and still get dirty, still be on the tools. When I took on more than he could actually run, I pulled him off tools to babysit a second crew. The math on both jobs went sideways simultaneously. Nobody noticed until the invoices went out wrong and the change orders that should have been written weren't. That's how a working foreman stops earning his keep, and it happens fast.
What Happens to Your Existing Jobs When You Take One Too Many
Here's the part nobody talks about. Overcommitment doesn't just slow down new work. It bleeds margin out of jobs already in progress.
The ways it happens are predictable. Punch-outs get rushed because the crew has somewhere else to be. Inspections get scrambled when someone calls in sick and the truck gets rerouted. A foreman covering two sites means the guys at site two aren't sure what they're doing for the first forty minutes of the morning. Change orders don't get written because everyone's moving.
The margin bleed almost never shows up on the job where you made the scheduling decision. It shows up on the jobs you already had in progress.
After the crash, I rebuilt on service-only with three men. Lean, slow, careful. And even then — one more call at 3:30 on a Thursday afternoon, said yes, and a three-man service operation starts cannibalizing itself inside a week if the dispatcher isn't treating the schedule as a hard ceiling. The shops that limped through 2009 weren't just undercapitalized. Some were also hemorrhaging cash through bad scheduling on work they were already committed to.
The tankless situation is a good example of the silent kind. When a crew is stretched, the gas-line upsize gets eyeballed instead of measured. The Navien goes in. Eighteen months later the callback comes — unit's short-cycling, homeowner's furious, warranty claim is gray because the gas pressure was never right. Nobody connects that callback to the scheduling decision made a year and a half ago. It just looks like a warranty issue. Gets eaten. The original job, already thin, just got thinner retroactively.
The Contrarian Take: Turning Down Work Isn't Discipline, It's Just Math
The trades coaching world has done a number on this subject. "Say no to the wrong work." "Protect your calendar." "Charge more, take less." Some of it is true. But the way it gets sold, turning down jobs sounds like a virtue. A sign of a serious shop. Something to be proud of.
It ain't. Not by itself.
Turning down work without knowing your actual throughput number is a panic move dressed up as confidence. If your existing jobs are bleeding margin because you're overextended, and you turn down a new call, you've just reduced revenue without fixing the underlying problem. You don't get credit for the job you didn't take while the jobs you did take are quietly going sideways.
I've said this before about flat-rate pricing books: shops that price off somebody else's spreadsheet don't own their own numbers. Same thing here. Shops that cap their intake based on whatever formula some guy selling a coaching program put in his workbook — built for a different market, a different crew, a different job mix — are doing the exact same thing. They borrowed the number. A borrowed capacity ceiling is worth what you paid for it.
The guys selling this stuff, with the case studies about shops that tripled revenue by "protecting their capacity" — they've never personally had to figure out which man to let go and then make that call. They've never watched receivables stretch from 45 to 110 days. What they're selling skips the step where you actually calculate what yes costs you. Without that, you're not making a financial decision. You're just feeling busy and calling it strategy.
How to Calculate Your Real Throughput Ceiling Before Next Monday
This doesn't require software. Thirty minutes and a piece of paper.
Start with billable hours. One crew — working foreman and two guys, standard residential service unit. Forty hours a week on paper. Drive time, supply runs, callbacks, scheduling gaps eat into that. You're lucky to get twenty-six to twenty-eight billable hours out of that crew in a normal week. Use twenty-six.
Multiply by your blended billing rate for that crew's actual job mix. Not your best rate. The blended rate — some service calls, some installs, whatever you actually run. That's your weekly revenue ceiling for one crew at full efficiency.
Now be honest about the drag. If your foreman is also doing his own scheduling, calling suppliers, handling customer calls, and covering for the guy who runs late twice a week, you're not getting twenty-six billable hours. You're getting twenty-two. Maybe twenty. Use that number.
When I started in 2005, I knew this per truck. When I went to six trucks, I started looking at total volume instead. That's where I lost the thread. Total volume can look fine while individual trucks are underwater, and you won't know until you try to figure out where the margin went.
One calibration check before any of this matters: the 90-day question. Can you pay your guys, your truck notes, your rent, and your insurance for 90 days with zero new revenue? If the answer is no, this isn't a capacity management question yet. It's a survival question. Get cash right first. Capacity is second.
The Story: Spring 2008, Six Trucks, and the Jobs That Were Quietly on Fire
I can see it now more clearly than I could then. Spring and summer of 2008, Smith Mechanical was running behind on multiple jobs simultaneously. Not catastrophically — not the kind where someone calls and threatens to pull your contract. Just quietly behind. The punch-out list on a new construction job kept sliding. The foreman was making judgment calls on rough-in work that should have been coordinated with the GC. Change orders were written on scraps and turned in late.
I lost two builder accounts in the same week in October 2008. I told myself it was the crash. And the crash was real. But those relationships had been deteriorating for months before Lehman went down. The work was slipping. The crew was stretched. The GCs were already half-frustrated. The crash gave everyone an excuse to make the calls they'd already been considering.
The four men I laid off — I knew their names, I knew their trucks, I knew roughly what their situations were at home. One of them had a kid who'd just started high school. I'd hired him out of a shop in Shrewsbury that had closed. Running twelve guys at the production of eight is how I ended up making that phone call. Nobody forced me into it. I took too many jobs, I didn't track the margin on each one, and I found out the real cost when I had to do the math on who to keep.
That's not a story I enjoy telling. I tell it because nobody told it to me when I was running two trucks out of my garage and thinking about going to four.
What You Actually Do Monday Morning
Pull last quarter's completed jobs. Look at estimated hours versus actual hours on each one. Add up the gap across the quarter. That dollar figure is what overextension cost you — the hours you ate, the callbacks you absorbed, the punch-out time you didn't bill because you wanted the job closed.
If you don't track hours by job, that's the first problem. Fix it. Start this week with a lined notebook and a foreman with a pencil. Date, job, crew, hours in, hours out. You don't need an app. You need the data. Get thirty days of it and you'll already know more about your actual ceiling than you know right now.
Second: write a number on the whiteboard. What does this crew bill per day at full efficiency? When a new call comes in and that number is already hit — what's the answer? That's the policy. A number, written down, with a decision attached to it. Not a gut call at 3:30.
If you're running under five trucks and paying for software to manage capacity and you're still overbooked, the software isn't the problem. No scheduling tool fixes a shop that hasn't defined its own ceiling. It just makes the overbooking more organized.
The ceiling is a number. Go find yours.
FAQ
If I'm booked out six weeks, doesn't that mean I should be hiring, not turning down work?
Maybe. But six weeks booked means nothing if you don't know whether those jobs are priced right and whether your current crew can actually make money on them. I've seen shops with eight-week backlogs bleeding money on every job because they were already understaffed on work they had. Hire when your margin is holding and you've genuinely run out of crew. Don't hire to fill a backlog you haven't checked. Fixed cost hits immediately.
How do I tell a long-term customer I can't take their job right now without losing them permanently?
Call them. Don't email. Tell them straight: you're committed to doing the job right and your crew is full through a specific date, but you want their work and you'll hold a slot. Give them an actual date. Not "a few weeks." Most customers who've worked with you will wait if you're honest. What loses them isn't the wait. It's "we'll try to fit you in" turning into three weeks of nobody calling back.
What's the actual difference between a shop that's productively busy and one that's overextended?
From the outside, nothing. Full schedules, trucks on the road. The difference shows up in job costing. A shop running right knows what each job made when it closes. An overextended shop finds out at the end of the quarter that margin wasn't what the estimates said, and nobody's sure where it went. If you have to wait for the quarter to close to answer that question, there's your answer.
Is there a rule of thumb for how many jobs one working foreman can actually run before the margin drops?
A working foreman splitting days across more than two active sites is already in trouble. The moment he stops being on the tools and starts being a coordinator, you've lost the efficiency you were paying for. One site, full presence, with a second he can check at end of day — that's the ceiling. I've never seen the math work past two.
My slow season always follows my busy season — won't turning down spring work make the fall cash problem worse?
Yes. Turning down spring work without a plan makes fall worse. That's why your capacity ceiling and your 90-day cash question have to be answered together. If you don't have 90 days of operating expenses in reserve, the priority is building that buffer — which means maximizing margin on what you take now, not volume. Price the spring work right. Close the invoices fast. That builds the fall cushion. Taking every call at thin margin doesn't.
How do I know if the margin bleed is from overcommitment or just bad estimating?
Look at where the hours went. Bad estimating shows up on specific job types — you consistently underestimate rough-in on a certain house type, or the fixture allowance is always wrong. Overcommitment shows up across everything as overtime, callbacks, and punch-out that drags past schedule. Hours blowing on every job type regardless of scope — that's overcommitment. Concentrated on certain scopes — that's estimating. Different problems, different fixes.
Enjoyed this article?
Get articles like this in your inbox every Monday. Free, no spam.