Subcontractor Utilisation and the Capacity Question Behind It

Why this matters

Subcontractor utilisation gets filed next to the cost figures and read as one, and it is not a cost number at all: it holds no money. It is the only figure in the shop that says what share of the work you took on was work you could not do yourself. Read as a standing measure of how much capacity you are renting, it answers a worker-classification question, a hiring question, an insurance question and a quality question nothing else in your records asks. A shop that lets it drift upward without noticing ends up with a crew it never hired, a warranty book full of work nobody from the shop inspected, and no idea when either started.

What it counts

The share of jobs completed in the window that carried at least one subcontractor. It is a binary per job: the job either had one or it did not.

Four flattenings follow, and each changes how you read a move:

• A job with four subcontractors counts the same as a job where a helper came for half a day. A fully brokered project and a single trade filling one gap are one value here.
• There is no money in it. It is a count share and must never be set beside a cost share and read as the same quantity. A quarter where subcontracted jobs are 25 percent of the job count and 55 percent of direct cost is completely ordinary, and treating the two as one number is the most common way this figure gets misquoted.
• The denominator is every completed job, including the free callback and the one-hour service call. Run a quarter heavy on small jobs and the same subcontracted work reads as a lower share.
• It counts jobs that CARRIED a subcontractor, not jobs that needed one. An assignment made and then stood down still counts if the row stayed on the job.

One more, shared with its siblings: the window selects jobs, so a recent window is not finished - a job opened last month can still pick up its first subcontractor assignment next month, and a fresh quarter read against a matured one reads low. The companion card in the references works this through.

The worked quarter

A shop completed 86 jobs. Twenty-two carried at least one subcontractor, so utilisation is 22 over 86, or 25.6 percent of completed jobs.

Index subcontractor cost in sub-days: 1.0 unit is one subcontracted worker on site for one working day. Those 22 jobs carried 41.0 units between them.

The flattening in figures. Of the 22 jobs:

• 14 carried one subcontractor for part of a day
• 5 carried one subcontractor across multi-day scope
• 3 carried three or more subcontractors each

Those 3 jobs hold 21.0 of the 41.0 units, 51.2 percent of the quarter's subcontractor cost, while being 3 of 86 completed jobs, 3.5 percent of the job count. Two bases, two quantities, and the headline share treats all 22 jobs identically.

Against the same quarter last year, 18 of 79 completed jobs or 22.8 percent, utilisation is up 2.8 percentage points. That is the number most shops look at, and on its own it is not worth a meeting.

The cut that tells a strategy from a drift

Sort the 22 jobs by the reason each one needed a subcontractor, and ask two things of each reason in this order: is the work lawfully buyable from a subcontractor at all, and could this shop plausibly hire for it? The second is arithmetic. The first is not, and it has to be settled before the arithmetic runs.

• 12 jobs: overflow in the shop's own trade at peak, work the crew does every week and simply could not reach. Hireable - and check first whether it is lawfully buyable at all, because a worker performing the core service you sell, on your jobs, to your schedule, week after week, is the paradigm case of a misclassified employee under the IRS common-law control test and the FLSA economic-reality test, and in an ABC-test state such as California (Labor Code 2775), Massachusetts or New Jersey it fails outright on the prong requiring the work to sit outside your usual course of business. Put this cut in front of your own accountant or employment attorney before you size it: what is at stake is back wages, payroll tax and comp premium, not a margin point.
• 7 jobs: a licensed trade the shop does not hold and has no intention of holding. Permanent, correct to buy.
• 3 jobs: specialist one-offs, a lift, an engineering sign-off, a remediation scope. Permanent, correct to buy.

The hireable count is what carries the decision: 12 of 86 completed jobs, or 14.0 percent. Last year that cut was 7 of 79 jobs, or 8.9 percent, so it rose 5.1 percentage points of the job count while the headline rose 2.8 points of the same base - a little over half the movement, in the figure that actually decides something.

Then size it in days, because a hire is a days question. Those 12 hireable jobs consumed 26.0 sub-days in the quarter, so about 104 sub-days over a year at that rate. The obvious denominator is a technician's 240 working days a year, 260 weekdays less about 20 for holidays and leave, and it is the wrong one. Those two sides are not the same kind of day: 104 is subcontracted days delivered on site, while 240 is a technician's gross paid working days, out of which shop time, travel, training and non-productive hours still have to come. Take your own measured on-site days per technician per year and divide by that instead. It will be materially under 240, which puts the overflow habit at a higher share of a technician's deliverable year and brings the hire threshold down with it.

At an illustrative 180 on-site days, the figure to replace with yours first, 104 over 180 is 58 percent of one technician's deliverable year rather than the flattering 43 percent the gross base reports. Real, growing, and still not a hire on volume alone. It becomes one at 180 sub-days a year, 45 in a quarter against the current 26.0, so the overflow work has to grow by about 1.7 times rather than the 2.3 the gross base implied. Either it grows to roughly that or it stays a buy, and anything between is a service-level decision, not arithmetic.

Bands, and the honest thing to say about benchmarks

There is no defensible industry benchmark for subcontractor utilisation, and anyone quoting one has invented it: trade mix, licence scope and market make a residential service shop and a light commercial installer incomparable here. Benchmark against your own trailing history instead - four to eight quarters of your own hireable cut is the only real answer available.

These bands read the hireable share, not the headline, for the reason the closing cases make plain:

The worked shop sits at 14.0 percent, one point below the top of the standing-habit band and rising 5.1 points year on year: sub-day test every quarter, and probably the next band inside a year.

Reading a move before you act on it

A move in this figure does not explain itself, and several things that move it are not capacity decisions. The last column is the part to use.

The expensive version is hiring against a leave cover: the hireable cut jumps for two quarters, the shop reads growth, and the new technician arrives the quarter the original one comes back.

What rises with the share, whether or not cost does

Your name is on the invoice, the warranty and the callback, whoever did the work, so every point this figure rises is a point of your finished work somebody outside your shop produced. Three exposures rise with it:

• Work nobody from the shop looked at. The control is cheap and specific: a named person from your shop inspects subcontracted scope before the job closes, signed off on the job under their name. Without that, the first inspection is the customer's.
• Insurance that was current at onboarding and is not current today. A certificate has to be in force on the day of the work, which means expiry tracking rather than a folder - and it has to show workers compensation as well as general liability, because in most states an uninsured subcontractor's workers fall back onto the hiring contractor's own policy as a statutory employer, and your comp auditor will charge premium on every dollar paid to a sub whose comp certificate you cannot produce at audit.
• Callbacks you cannot attribute. If your callback records do not say whether the fault was on subcontracted or own scope, you cannot tell whether the rising share is costing you anything. Record the attribution from the first subbed job, not from the day it becomes a problem.

Hold-backs and retainage are the other lever people reach for. What a hold-back actually entitles you to is a question for your own attorney and turns on the agreement you signed, so get it right in the agreement rather than inventing a remedy after a defect.

Two shops at the same share

Both sit at about 26 percent, and neither is the worked shop above.

Shop A: 23 of 88 completed jobs carried a subcontractor, 26.1 percent, but 20 of those 23 were a licensed trade it does not hold and never will, so its hireable share is 3 of 88 jobs, 3.4 percent. Nothing here is drift, and its capacity question is answered: its work is vetting, insurance expiry and inspection sign-off.

Shop B: 21 of 81 completed jobs carried a subcontractor, 25.9 percent. Nineteen of those 21 were its own trade at peak. Its hireable share is 19 of 81 jobs, 23.5 percent - deep into the band where a shop has to write down whether it means it. This shop is renting a crew, quarter after quarter, and calling it a cost line - and renting a crew in your own trade at 23.5 percent of completed jobs is precisely the fact pattern a wage-and-hour or workers compensation auditor reads as employment, so its first move is a classification review with its own attorney, not a band on a chart.

The headline cannot tell these two apart, and no band applied to it can. That is why the bands above read the hireable cut, and why both questions have to be attached to this figure before it means anything.

References

• See related: Is This Worker a 1099 or a W2? Decision Tree - the classification gate that runs before the sizing
• See related: The Subcontractor Onboarding SOP - where the certificate and agreement checks live
• See related: Subcontract vs Employee vs Temp Decision Matrix - the decision this figure feeds
• See related: Vetting a Subcontractor Before You Put Them on a Job
• See related: COI Tracking for Subcontractors - certificates current on the day of the work
• See related: Subcontractor Spend Is Anchored to When the Job Was Created - owner of the job-anchor claim
• See related: Who Is Liable When a Subcontractor Causes Damage