Your own drivers or a courier service? Run the numbers

Every local business that delivers reaches this decision, usually twice: once when demand outgrows the owner’s car, and again when the courier invoices start looking uncomfortable.

The comparison is almost always done badly, in both directions. Courier fees get compared to a driver’s hourly wage, which understates the driver. Or in-house gets defended on cost per delivery on the busiest day of the year, which understates the courier. Here is how to do it honestly.

Load your own cost properly

A driver’s wage is roughly half of what a driver costs. The full picture, per driver, per month:

• Wages, including overtime and the hours around the route — loading, returns, paperwork.
• Payroll costs on top: employer taxes, workers’ comp, benefits if you offer them, paid time off.
• The vehicle: lease or depreciation, insurance, maintenance, tires, tax, parking.
• Fuel, at real consumption on your actual routes, not the manufacturer’s figure.
• Idle time: the hours paid when volume is thin. This is the number people leave out, and on a quiet Tuesday it is most of the cost.
• Management: the dispatcher’s time, HR, recruiting, training a replacement twice a year.
• Software and phones.

Add it up per month, divide by deliveries completed that month, and you have a real cost per delivery. Do it for a busy month and a quiet month separately — the two numbers will be far apart, and that gap is the whole argument.

Load the courier cost properly too

The fee is not the cost either.

• The fee itself, plus surge or peak multipliers, which apply exactly when you need capacity most.
• Tips, if the model expects them.
• Failed and returned deliveries, and who eats them.
• Support time: someone in your office chasing a lost order is your labor cost, not theirs.
• Refunds and goodwill for deliveries you could not control.

Then the parts with no invoice line, which are the ones that actually decide this for many businesses:

• You do not own the doorstep. The last human your customer meets does not work for you and is not trained by you.
• You do not own the data. Which addresses are slow, which are wrong, what your real service times are — you get little of it.
• You do not own the brand. The tracking page, the notification, and the person at the door are all somebody else’s.
• You do not own the capacity. On the busiest day of your year, you are bidding for drivers against everyone else in town.

None of that is an argument against couriers. It is an argument for pricing them at what they really cost.

Where the break-even usually sits

The variable that decides it is density: deliveries per driver-hour within your radius.

Courier per-delivery pricing is close to flat. Your own cost per delivery falls steeply with volume, because the expensive parts — the van, the insurance, the paid hour — are fixed once the driver is out. Two deliveries an hour is punishing. Four is comfortable. Six changes the business.

So the shape is simple: below some volume, couriers win on cost and on flexibility; above it, your own drivers win, and the gap widens with every additional stop per hour. Work out where your own line crosses the courier’s flat one, and you have the answer for your business rather than someone else’s.

The number that moves the crossing point is not the wage — it is stops per hour. A fleet that improves density from three to five stops an hour cuts its own cost per delivery by roughly a third without touching pay. That is why planning and routing show up in a build-versus-buy decision at all.

The hybrid nearly everyone lands on

Most fleets that grow past the decision do not pick one. They run their own drivers for the predictable base load and use couriers for the parts that would otherwise force them to hire:

• Peaks. December, the holiday weekend, the day the promotion lands.
• The long tail. The one delivery forty minutes outside your radius that would cost a driver ninety minutes round trip.
• Overflow. The last six orders on a day that ran long.
• New territory, before you know whether the volume is real.

This is the structurally right answer more often than either extreme. Your own drivers carry the volume you can predict, at your cost, under your brand. Couriers absorb the variance you would otherwise have to buy a van for.

It only works if you can see both in one place — which is the practical requirement most fleets discover late.

When to keep it in-house regardless of the math

Some deliveries are not really about cost per delivery.

If the goods are controlled, age-restricted or high-value, you need enforced proof and a person you employ. If installation, assembly or a returned item is part of the job, a courier is the wrong instrument. If your delivery experience is genuinely part of your product — and for some businesses it plainly is — then handing the doorstep to a stranger is a strategic decision, not a cost one.

Conversely, if delivery is a cost center you would rather not think about, and volume is thin and spiky, paying somebody per delivery to make it disappear is a perfectly good answer.

The one-page exercise

Take last month.

1. Total deliveries completed.
2. Your fully loaded in-house cost for the month, including idle time and management. Divide by deliveries.
3. What the same deliveries would have cost at courier rates, including peak multipliers on the days you would have hit them.
4. Repeat both for your quietest month and your busiest.
5. Compute stops per driver-hour. Ask what that number would be with better planning, and rerun step two at the improved density.

Four numbers and one honest question. Most owners find the answer is not the one they have been assuming, and that it is different for the base load and for the peak.

PrivateMile gives you the in-house side of that math from work actually completed — cost per delivery, stops per driver-hour, fuel and distance per vehicle, driver pay from real routes — rather than from estimates, and its plans price by order volume with drivers unlimited, so adding a seasonal driver does not add to the bill.