Buy or Rent: How an Excavation Contractor Can Tell Which Machines Are Worth Owning

The problem: Rental invoices are visible every month while the cost of owning a machine is scattered across payments, insurance, repairs, and a yard nobody bills, so the buy-or-rent decision gets made on the visible half.

The solution: Track hours by machine against total cost of ownership, so you can see the break-even point each class of equipment actually has to clear.

Every excavation contractor has this argument, and it never gets resolved, because both sides are arguing from the number they can see.

The operations side sees rental invoices. A mid-size excavator on a six-week dig, another one the following month, a compactor for three weeks, and by October the rental line looks like a down payment on a machine the company could have owned outright. "We are renting the same machine four times a year. We should just buy it."

The finance side sees the payment, the insurance, and the last time a machine sat in the yard through a slow February. Neither side is wrong. They are just comparing a number that arrives on an invoice against a number that is spread across six accounts and a piece of gravel behind the shop.

Utilization is the entire question

Strip everything else away and buy-or-rent is one question: how many hours a year will this machine actually work.

Above a certain number of hours, owning is cheaper per hour than renting, and the gap widens fast. Below it, renting is cheaper, and the gap widens fast in the other direction. There is a break-even, it is specific to the machine class and to what you can get it for, and it is not hard to calculate. The problem is not the math. The problem is that almost nobody at this size can tell you how many hours the machine they already own worked last year, so they have no basis for predicting the one they are considering.

Ask most contractors and you get an answer shaped like "it stays pretty busy." Which is true, and useless. A machine that runs 1,400 hours and a machine that runs 550 hours both look busy from the office, because both of them are out on a job most weeks. The difference between them is whether the purchase was one of the best decisions the company made that year or one of the worst.

What ownership actually costs

The other half of the trouble is that the cost of ownership never appears anywhere as a single number.

The payment is in one place. Insurance is in a policy covering the whole fleet. Repairs are in a maintenance account with everything else. The mechanic's time is in payroll. Transport between jobs is in a lowboy cost nobody assigns. Depreciation is a year-end entry from the accountant. The yard space is not charged to anything at all.

So when the owner compares "renting this cost us $38k last year" to "the payment would be $2,400 a month," he is comparing a complete number to a partial one. The rental figure includes maintenance, insurance, transport in some cases, and the certainty that when you do not need it, you stop paying for it. The ownership figure he is holding includes none of that. Almost every buy-or-rent decision at this size is made against that mismatch, and the mismatch reliably points toward buying.

Both directions of being wrong are expensive

It is worth being clear that this cuts both ways, because the usual framing treats renting as the safe choice.

Buying a machine you use 500 hours a year means paying full ownership cost for roughly a third of a machine's working capacity. On a piece of equipment costing $95k a year all-in, that is real money that bought idle time.

Renting a machine you use 1,400 hours a year is the opposite error and is often larger. At typical monthly rates, a machine kept on rent nearly year-round can cost well over twice what owning it would, and because the invoices come in monthly and get coded to job cost, they never accumulate into a number that prompts anyone to look. The company that has been renting the same class of machine every month for three years is usually making a bigger mistake than the company that bought one too early.

How the answer becomes available

The missing piece is hours by machine, and the machines are mostly already telling you. Telematics on newer equipment reports engine hours. Older units have a meter someone can read. Operators already fill in daily reports.

The change is that the hours land against the machine and the job, rather than dissolving into a daily report that gets filed. Once hours sit alongside the payment, the insurance allocation, the repair invoices coded to that unit, and the transport, you have a real cost per hour for every machine you own. That number is what you compare a rental rate to, and it is the only honest version of the comparison.

Then it should keep itself current. Instead of someone rebuilding an equipment analysis before each purchase decision, automation totals hours as they come in and flags the two conditions that matter: an owned machine tracking below the hours that justify it, and a rented class of machine crossing the hours where owning would have been cheaper. The decision stops being an argument and becomes a report that is already sitting there when the question comes up.

A look at an excavation contractor

Take an excavation contractor doing about $11 million a year, running site work and utilities, with a mix of owned machines and a standing relationship with two rental houses. The rental line last year was around $310k, which the owner has started calling "a machine and a half we could own." He is close to buying two excavators.

Suppose the company reads engine hours by unit for twelve months and pulls together what each owned machine truly costs: payment or depreciation, insurance, repairs coded to the unit, mechanic hours, and transport. You would expect a mixed picture rather than a clean answer.

Say a mid-size excavator works out to roughly $95k a year to own, all in. Against the rate the company pays for the same class, break-even lands somewhere near 900 hours. One machine the company already owns runs about 1,300 hours, and owning it is clearly right. Another runs about 500, which means the company is spending on the order of $40k a year on capacity it does not use, and that machine would likely be better sold and rented back on the weeks it is needed.

For the two purchases on the table, the same exercise gives an actual answer rather than a preference. If the rental history shows one class of machine on rent for 1,100 hours across the year, buying it is likely to pay. If the other class only reached 400 hours and the peaks all fell in the same two months, buying it would be paying year-round for a seasonal need. The contractor who has this in front of him buys one machine instead of two, and is probably better off for it.

How to start

You can get to a defensible answer within a year, using equipment you already have.

1. Capture hours by machine. Telematics where it exists, a meter reading on a weekly report where it does not.
2. Build one true cost per owned machine. Payment or depreciation, insurance, repairs, mechanic time, transport. One number per unit per year.
3. Calculate the break-even hours for each class. Divide annual ownership cost by the rental rate you actually pay. That is the hurdle.
4. Check your rental history against it. Any class you rented past its break-even hours is a buy candidate. Any machine you own below it is a sell candidate.
5. Let it stay current. Set automation to total hours and flag both conditions as they happen, so the next purchase question is already answered.

The takeaway

Buy-or-rent is not a philosophy, it is an hours question with a break-even you can calculate. What makes it feel like a philosophy is that rental invoices are visible and ownership costs are scattered, so the two sides of the comparison are never the same shape. Get hours onto the machine, get every cost of ownership onto the same unit, and work out the break-even for each class you use. You will probably find you own something you should not and rent something you should own, and both of those are worth knowing before you sign for the next machine.