In-house shop or outsourced maintenance?

The comparison almost every fleet makes first: the vendor charges $150 an hour, our technician makes $34, so in-house saves $116. Everyone who has run a shop knows that is wrong. Worked properly, on the same two numbers, the gap is closer to $48.

The right comparison is the vendor’s invoice against your fully loaded shop rate — what one hour of productive wrench time costs once you count everything that has to be true for that hour to happen.

What goes into a loaded rate

Wages including overtime and callout. Benefits — your finance office almost certainly publishes an approved fringe rate already, and public-sector loads run high, which is part of why the wage-to-invoice comparison flatters in-house so badly. Training and certification, including the time away from the bay. Tooling, lifts, scan tools and OEM diagnostic subscriptions, which renew annually and once per make. Facility, utilities, waste fluid disposal and parts room space. Supervision and administration — the shop supervisor and whoever closes work orders are real cost with no wrench time.

Then the one that decides the answer: non-wrench time. It belongs in the denominator.

Wrench time is the divisor

A technician is paid for 2,080 hours, is not available for 2,080, and is not turning wrenches for however many they are available. Chattanooga’s internal audit used a clean version: start at paid hours, deduct leave, holidays, sick and training, and you reach roughly 1,760 hours. The method travels better than the number — run your own actuals through it.

Then take another cut. Meetings, road calls, parts chasing, cleanup, waiting. The share that lands on a work order is your direct labour percentage.

A $34/hour technician at 36% fringe costs about $96,000 in wages and benefits. Add $30,000 for that technician’s share of supervision, facility, tooling and administration and you are at $126,000. Divide by 1,760 hours at 70% direct labour — 1,232 billable hours — and the loaded rate is about $102 an hour.

Every input there is a placeholder, including the 70% the callout tells you not to assume. Hold everything else and drop direct labour to 55% and the same shop costs $130 an hour: the advantage over a $150 vendor falls from about a third to about 13%, inside the range that vendor turnaround or parts pricing can erase.

Springfield, Missouri publishes real numbers against this: a policy of holding internal rates within 77.5% to 82.5% of average private sector market rates, with FY2023-24 actuals of $127.11 light duty and $157.43 heavy. That band is a pricing policy as much as a cost finding, and the budget book never says who runs the survey — so it tells you a well-run public shop can be modestly cheaper than its market, not that the market is four times more expensive.

The break-even question

Below some fleet size you cannot keep a technician busy, and a technician idle half the time has a loaded rate double the one you calculated.

The usual method converts the fleet to vehicle equivalent units — a car is 1.0, a pickup roughly 1.5, fire apparatus around 10. Chattanooga used about 120 VEU per mechanic. Know where that came from before quoting it at a council: the audit’s own methodology credits it to two consultancies and to Government Fleet magazine, with no sample disclosed anywhere in the chain.

Carried out to one, two and three technicians it gives a rough screen — under ~100 VEU one full-time technician is hard to justify; 100–250 is the difficult range where any absence puts you back with a vendor; above ~250 you can staff to a schedule. Those thresholds are not a published standard, and composition moves them a long way: low-mileage sedans generate far less work per VEU than a refuse fleet on the same count. Use the screen to decide whether the question deserves a full study, not to answer it.

The hybrid model

Almost nobody runs a pure model. Fleets that call themselves “in-house” are usually running a hybrid with no written policy for it.

The split that works: keep PM, inspection and light repair in-house, because that is predictable, schedulable work that fills your bays and gives you control of availability. Outsource heavy engine and transmission work, body and paint, glass, towing, tires, aerial certification, and specialised systems where tooling or certification cost cannot be spread over your volume.

The failure mode is not the split — it is doing it without criteria. Denver’s 2024 audit found the city had spent nearly $5 million on about 800 outsourced jobs across 2022 and 2023 without documented policies, criteria, or cost-benefit analysis. Staff could name the categories that always go out, but said the main reason for everything else was understaffing, and that reasoning was never written down or costed.

Warranty and recall

You do not lose warranty coverage by servicing in-house. Under 15 U.S.C. 2302(c) no warrantor of a consumer product may condition its warranty on using a branded article or service without an FTC waiver.

The part fleets get wrong is which units that covers. Coverage turns on product type, not who bought it: 16 CFR 700.1(a) says the use an individual buyer puts a product to is not determinative, uses automobiles as its worked example, and resolves ambiguity in favour of coverage. A city-owned sedan or half-ton pickup is normally still a consumer product, city plates or not. The Act stops at equipment not normally used personally — fire apparatus, refuse packers, Class 8 tractors — where your protection is the written OEM warranty, which is contract rather than statute. Either way, keep records: a denied claim turns on proof of interval-compliant service.

Recall work is free, and usually not yours to do. Under 49 U.S.C. 30120(a) the manufacturer must remedy a safety defect without charge when the vehicle is presented, choosing repair, replacement or refund; in practice it runs through a dealer. The free remedy lapses 15 calendar years after first purchase — five for a tire. Inside those windows, an open-recall repair your shop performs is money the manufacturer was required to spend. Run VIN-level recall checks on a schedule, not when a letter arrives.

The arguments that survive when outsourcing is cheaper

Control of scheduling. In your own shop you set the queue. With a vendor, PM slots slip first because they are the appointment with no complaining operator behind them. A deferred PM costs whatever it was meant to catch plus the unplanned downtime around it — see PM compliance.

Emergency response. Snow events, a plow down at 3am. Vendors price this as a premium call-out if they cover it at all. Where your service obligations include response times, that capability is why the shop exists.

Institutional knowledge. A technician who has run the same twelve refuse packers for six years diagnoses faster and knows which units are worth repairing. It does not come back once you disband the shop, and rehiring means competing in a tight market — Denver’s staff told auditors they lose technicians to private-sector compensation, and the division resorted to hiring bonuses.

Availability. Denver targets 95%; across nine quarters to March 2024 it reported 90–94% and never reached the goal, running 54 technicians against a consultant’s recommended 76. Availability is what your customer departments actually feel, and it moves with bay capacity and queue priority — both of which you set in your own shop and negotiate in someone else’s.

None of this licenses skipping the arithmetic. Do the loaded rate, state the gap plainly, then argue the strategic case on its own terms. “We are 15% more expensive and here is what that buys you” is far more credible than a savings claim that dies under a calculator.

The compliance obligation does not transfer

If your fleet is subject to federal motor carrier rules, outsourcing changes who turns the wrench and nothing else. 49 CFR 396.3 requires the carrier to inspect, repair and maintain — “or cause to be” — and to hold a per-vehicle maintenance record. That is a per-vehicle file, not a shoebox of receipts; your vendor holding your invoices is not you holding your records. On brakes, 396.25(a) makes the carrier responsible whoever performs the work, while the qualification-evidence duty in 396.25(e) is written around employees and does not by its terms reach a vendor’s technicians — so with an outside shop that evidence has to come from your contract rather than the regulation.