Service Work vs. Install Work: You Can't Run Both the Same Way
Service Work vs. Install Work: You Can't Run Both the Same Way
In 2008 I had six trucks on the road. Two builder accounts. A decent residential service book. Guys I'd hired personally, trained personally, counted on. By November I'd laid four of them off. Not because the work disappeared overnight. Because the money stopped moving and I'd built a shop that couldn't tell the difference between revenue and cash.
It doesn't announce itself. That's the problem.
The Drift Problem: How Most Shops End Up Doing Both
Smith Mechanical started in my garage in 2005 as pure residential service. Hot water heaters, boiler calls, drain work, fixture swaps. One truck, one phone, and Mike Callahan from my Beacon days who needed hours. That was it.
Then a builder called. Then another. By 2007 I had four trucks doing both — service calls in the morning, rough-in work for two residential builders in the afternoon, sometimes the same truck doing both in the same day. By early 2008 I had six trucks and what I thought was a thriving shop.
What I actually had was two businesses running on the same schedule, the same crew, and the same pricing instincts.
I was counting success in trucks. The builder work had long collection cycles. The service work had thin margins because I was treating it like an afterthought. When both were running full speed, the numbers blurred into something that looked like profit. Then October 2008 happened and the blur cleared up fast.
Why Service Work and Install Work Are Actually Two Different Businesses
They share a truck and a set of tools. That's about it.
A service call needs a tech who can walk into an unknown situation, figure out what's wrong, tell the homeowner what it costs, and fix it — all in the same visit. Your truck has to be stocked to close without a parts run. Your rate has to absorb the diagnostic dead-ends, the windshield time, the callback you'll eat twice a year, and the two-hour job that billed as forty-five minutes because the last guy installed it wrong. None of that math works on a five-fixture rough-in.
An install job runs on lead time, material staging, and sequencing. You need someone who can move three guys through a job in order — rough-in, inspection, trim-out — without losing a day to disorganization. That's the working foreman. The guy who still gets dirty but keeps the crew moving. I wrote about this for a regional trade rag in 2014 and I'm still mad about it. Shops stop treating that role like a real job and wonder why install crews cost so much to run.
When you lose your working foreman, you either overpay a project manager who won't touch a tool, or you underpay a lead tech who can't actually manage a job. You're paying full freight either way and getting half the output.
Then there's cash. Service work, priced right and collected at completion, pays COD or net-30. Builder install work pays in sixty days if you're lucky, ninety if you're normal, and never if the builder's bank pulled the construction line in Q4. A shop floating both on the same payroll and the same truck notes is the shop that runs out of money in month four of a slow service season.
What Happened When Smith Mechanical Tried to Run Both at Full Speed
By spring 2008 I had two primary builder accounts — solid volume, felt like a real business relationship. I'd let the residential service side drift. Not collapse. Drift. My best diagnostic tech was spending three days a week on rough-in work because that's where the trucks were pointed.
October 2008. Two builder accounts go quiet the same week. No phone call. Jobs slowed down. Calls stopped getting returned. The next phase kept getting pushed. Within three weeks my receivables had stretched from forty-five days to over a hundred. A pile of money sitting out there I couldn't collect and couldn't lien fast enough to matter in the short term.
The service side should have covered the gap. It didn't. I didn't have the call volume. The customers from 2006 had found someone else. My best diagnostic tech had been running rough-in for eight months and wasn't closing at the door the way he used to. You stop using a skill long enough, it goes soft.
I laid off four men in November. Kept three. Ran service-only through most of 2009. Built back slowly, and I stayed away from builder dependency for years after — not because I was being smart about it, because I was scared. But the lesson wasn't that builder work is bad. The lesson was that I'd built a mixed model by accident and had no system underneath either half. When one side collapsed, I had nothing to catch me.
The Contrarian Take: You Can Run Both — But Only If You Decide To
Here's where I'll push back on the "just niche down" crowd.
Most coaches selling that line have never had a February in Worcester where new install work dries up for six weeks and a boiler service call is the only thing keeping payroll covered. In New England, a pure install shop prays for mild winters and construction financing that doesn't fall apart in Q4. I've seen that prayer go unanswered. Hell, I lived it.
You can run both. I've watched shops come out of 2009 doing exactly that. The ones that stayed healthy ran service and install as separate cost centers. Same crew sometimes, same trucks — but separate scheduling windows, separate markup logic, separate rules about what gets accepted and what gets turned down.
The killer isn't the mixed model. The killer is the accidental mixed model where every job gets priced the same way.
Flat-rate book pricing applied across both models is where this breaks down fastest. A flat-rate price built for a service call carries margin for windshield time, diagnostic work, callback risk, and the parts run you had to make. That math has no business on a four-day install. And a lump-sum install price has no room for the hour and a half a service call actually burns when the homeowner can't tell you what's wrong beyond "it's making a sound."
If you're pulling flat-rate prices from a subscription book for both models without rebuilding the math yourself, you don't own a pricing model. You own a guess someone else made about their shop.
How to Price Each Model Without Using Someone Else's Spreadsheet
The cost drivers aren't the same. That's the whole problem.
Service work: you're paying for a tech who can diagnose, not just install. You're paying for truck stock so you can close on the first visit. You're paying for same-day availability — that's a real premium and most shops don't charge for it. Dispatch overhead, callback risk, the call that ran three hours because the shutoff valve behind the wall didn't hold — all of it belongs in your service rate. Not eaten as overhead. In the rate.
Install work: you're floating material. You're paying for staging time, foreman coordination, the day lost when inspection fails. On builder work, you're carrying retainage — often ten percent held until punch-list closeout — and the punch list itself gets used as collection leverage. I know that one up close. I wrote about Whitman Builders before: I fired them in 2011 over slow pay. What I've said less is that by the time I pulled the plug, they were carrying serious AR with us and the balance I did eventually see came back at cents on the dollar. Part of that was their problem. Part of it was mine — I'd never built install margin to survive a slow or partial collection. The profit I thought I was making assumed full payment on the contract schedule. Nothing in my number absorbed the actual risk of who I was working for.
Build your rates from your own cost of doing business. Subscription flat-rate books are useful your first year — training wheels before you have enough history to build your own. After that, they're a way to stop thinking. Your labor burden isn't their labor burden. Your callback rate isn't theirs. The moment you stop building from your own numbers, you're pricing someone else's business.
Monday Morning: How to Figure Out Which Model You're Actually Running
Pull every job from the last twelve months. Sort them two columns: service and install.
For each column, calculate average days to payment and gross margin. Not revenue — gross margin. Labor and material out, what's left.
Most guys who do this find out they've been subsidizing one model with the other without knowing it. Service work looked thin but paid in thirty days. Install work looked fat but average collection stretched seventy-plus days and two jobs had callbacks that ate the margin anyway. Or the opposite — installs were solid but service calls were priced off a rate set in 2019 and never touched.
Then ask yourself the staffing question honestly. Your best tech — is he a diagnostician who closes at the door, or a foreman who moves a crew through a rough-in? You probably can't get both in one guy at thirty-two an hour. If you're scheduling him for both, you're getting neither at full value. That's not on him. That's a role decision you haven't made.
One decision, by end of this week. Not "pick one model forever" — the seasonality point is real and I've already made it. The decision is: pick a primary model, name it, write it down, and run every scheduling and pricing call through it for the next ninety days. Service-primary or install-primary. One drives the bus, one fills the gaps.
Ninety days of that discipline will show you where your margin actually is.
FAQ
I'm a one-truck shop doing both service and install. Is any of this worth thinking about at my size?
More so at your size, not less. One truck means your time is the only asset you've got. If you're pricing a service call the same way you're pricing a two-day install, one of them is wrong. Probably both. Pull your last ten jobs and look at what actually paid per hour. The answer usually points clearly in one direction.
What if my service work is mostly warranty and callback — is that worth tracking separately?
Track it separately, but don't count it as service revenue. Warranty and callback work is a cost of doing business on your installs. If you're doing a lot of it, you've got a quality or pricing problem on the install side. Shops that lump callbacks into service revenue look like they have a busier service department than they do. You want to know what the jobs you actually charged for made. Callbacks cloud that number.
My builder account pays slow but the volume is good. When does slow-pay become a real problem?
When it stretches past your cash runway. If you can pay your guys and your truck notes for ninety days with nothing new coming in, a sixty-day collection cycle is annoying but manageable. If your runway is thirty days, sixty-day collections are a crisis with a date on the calendar. The volume argument from builders is real — it's also the argument every contractor who ever got burned by a bankruptcy made right up until the phone stopped getting answered. Sixty days, I can work with. Ninety days, I want a conversation. Past ninety, I file the lien and stop working. That's not a threat. That's a policy.
Is there a truck count where it makes sense to specialize?
Wrong question. It's not about truck count. It's about whether you have a working foreman. If you've got someone who can run an install crew and still get dirty, you can run installs as a real model. If you don't, you're running installs as a series of jobs where you're the foreman, the estimator, and the one chasing the building department. The model question and the foreman question are the same question.
Can I use the same tech for both service and install without wrecking both?
You can, but schedule him in blocks. A tech doing two service calls in the morning then driving to a rough-in for the afternoon is losing transition time and focus on both ends. Most techs are better at one or the other. Know which one yours is. Schedule him toward his strength. Fill the other side with someone else, even if it's just a helper for the install work.
I've been running both models but never separated the pricing. Where do I start?
Start with your labor rate. Take your actual fully-loaded cost — wages, burden, truck cost per hour, insurance, your own time — and divide it by real billable hours. Not hoped-for hours. Real ones. That's your floor. Then put the model-specific costs on top: callback risk and diagnostic time for service, material float and retainage exposure for installs. Whatever number you get, it's probably higher than what you're charging. That gap is what you've been giving away.
Enjoyed this article?
Get articles like this in your inbox every Monday. Free, no spam.