At a small fleet, most of what driver turnover costs is the idle truck, not the recruiting ad. When one driver leaves a 10-truck fleet, 10% of your revenue capacity sits in the yard until the seat is filled, on top of recruiting, orientation, and a slow first month for the new hire. ATRI's 2026 cost report puts 2025 turnover at 32% for fleets under 26 trucks, against 70.7% at fleets over 1,000, so a small fleet's job is to protect an advantage it already has.

That advantage is about to get tested. The driver market is tightening again, and the driver you have is going to get more phone calls. This post covers how to put a real number on one lost driver, and the day-to-day friction that makes good drivers take those calls.

Why do small fleets keep drivers when mega carriers cannot?

The American Transportation Research Institute's Operational Costs of Trucking 2026, released July 15, 2026 and covering 2025 data, puts overall driver turnover at 44.2%, down from 48% a year earlier. FreightWaves reported the split by fleet size: 32% at fleets with fewer than 26 trucks, 70.7% at fleets with more than 1,000.

Fleet size (2025, ATRI)	Driver turnover
Fewer than 26 trucks	32%
All fleets	44.2%
More than 1,000 trucks	70.7%

The overall drop is simple: in a soft freight market, drivers had fewer places to go. The gap between small and large is more interesting. It is proximity. At a 12-truck outfit the owner knows every driver by name, the dispatcher picks up the phone, and a problem at a dock gets fixed the same afternoon. At a mega carrier, a driver is a number talking to a different voice every call.

That edge is a behavior, not a budget line. Which means it can be lost without anyone noticing, one bad week at a time.

How do you put a number on one lost driver?

ATRI does not publish a cost per departed driver, and the round numbers you see on vendor blogs rarely cite a primary source. The honest way is to run your own. Here is the frame, with an illustrative 10-truck dry van fleet. These are example numbers, not benchmarks. Plug in yours.

1. The idle truck. Say one truck runs about 2,300 loaded miles a week at an average of $2.10 per loaded mile. That is roughly $4,830 a week in linehaul revenue. If the seat sits empty for four weeks while you recruit, that is about $19,300 you did not bill.

You do save the fuel and the driver pay on that truck. You do not save the truck payment, the insurance, the plates and permits, or the trailer sitting next to it. ATRI puts the 2025 average cost to operate a truck at $2.336 per mile, up 3.4% and a record, and $1.854 per mile excluding fuel. A lot of that keeps running whether the wheels turn or not.

2. Recruiting. Job board spend, referral bonuses, and your own hours on the phone screening applicants. Count the hours. They are the hours you did not spend covering freight.

3. Orientation. Paid orientation days, the drug test, the MVR, getting the new driver set up on your equipment and your apps, and the ride-along if you do one.

4. The ramp. A new driver does not know your customers, your shippers' dock habits, or which receiver makes you wait. The first several loads usually run slower, and the dispatcher spends more time on that truck than on any other.

Add those four up and you will have a number specific to your fleet. For most small carriers, the first line dwarfs the other three. That matters because ATRI also reports that truckload and refrigerated operating margins stayed below 1.0% in 2025. On margins that thin, a truck parked for a month is not a rounding error. It can be the quarter.

Why is the driver market tightening now?

The soft market that kept drivers in their seats is changing. C.H. Robinson's September 3, 2026 truckload update says federal enforcement actions continue to reduce the pool of available drivers, even for carriers not directly affected by the tighter CDL rules.

The same update says carriers are keeping higher hiring standards and chasing experienced, reliable operators, and that recruiting remains hard despite bigger recruiting budgets, referral programs, and wage investments. It expects more driver pay increases across parts of the industry, alongside elevated tender rejections and growing carrier selectivity.

ATRI adds the pay side. Driver pay rose at sub-inflation rates in 2025 for the second year in a row, and carriers reported driver wage costs up 1.6% on average in the first two months of 2026.

Put those together. The experienced driver who stayed put for two years because there was nowhere better to go now has recruiters calling with a sign-on bonus. Your 32% turnover rate was earned in a soft market. It is about to be tested in a firmer one.

What makes a good driver leave a small fleet?

It is rarely just the pay. Pay is the reason they give, because it is the easy one to say. The reasons they actually start taking calls are usually smaller and more daily:

Getting called three times for the same BOL photo. The driver sent it. It went into a text thread with forty other pictures. Now dispatch cannot find it and the driver is on the phone at 9 PM.
Getting blamed three weeks later. The broker disputes the invoice over a blurry POD, and the conversation starts with "why didn't you take a better picture" instead of "the system should have caught that."
Sitting at a dock with no word on what is next. The load is delivered and the driver has no idea whether there is a reload, a deadhead, or a night at the truck stop.
Eating a lumper fee. The driver paid cash at the dock, the receipt got lost in the cab, and nobody can prove it now.
Detention that nobody can prove. The driver sat four hours past the appointment and the only record is their memory.
Arguing about a declined load. The driver said no for a real reason, and the conversation turns into a fight instead of a reassignment.

Every one of those is a paperwork or communication failure. And every one is something the fleet controls. You cannot always match a mega carrier's sign-on bonus. You can stop making your drivers do the office's job from the cab.

How does the driver-side workflow protect the 32%?

This is where the tools matter, and where Howdy Dispatch is built differently. Most dispatch software is a board for the office. The driver app is an afterthought, or a separate product from a separate vendor. Eleos, for example, is a driver app that fleets connect to their existing dispatch system. Howdy Dispatch is the dispatch board and the driver iOS app from the same place, so a fix on one side shows up on the other.

Here is what the driver side does today:

Push notification on assignment. The driver hears about the load from the app, not from a missed call.
A "your next load" card. When a driver has no active job but one is scheduled, the app shows what is coming. Less sitting at a dock wondering.
One-tap pickup and delivery, each with a photo, plus manifest photo capture. The paperwork attaches to the load it belongs to, not to a text thread.
Self-delete on a bad photo within 30 minutes. If the driver catches a bad shot, they can remove it and retake it without calling anyone.
An offline outbox. Status updates queue up in a dead zone and send when the signal comes back, so a driver in a rural pocket does not get the "did you pick up yet" call.
Lumper receipt capture. The driver photographs the receipt, enters the amount, and tags how it was paid. HQ sees a per-load lumper total. It records and totals, it does not invoice the broker.
A detention timer. Pickup and delivery wait time is tracked against the appointment window, with a grace period the company sets.
Structured decline reasons. When a driver turns down a load, they pick from eight reasons, like equipment mismatch, distance, pay, or unsafe. The dispatcher sees why and re-routes, instead of arguing.

On the office side, the dispatcher sees load status and the truck on a live map, so the where-are-you calls mostly stop. And loads get built fast in the first place: our AI dispatch platform built on both sides reads a broker's rate confirmation PDF and pre-fills the customer, origin, destination, mileage, rate, and appointment times for the dispatcher to review and save.

None of this is a retention guarantee, and we do not have retention data of our own to quote. What it does is remove specific, daily frictions drivers complain about. That is the part of retention a small fleet can actually control this week.

One more thing to be clear about. Howdy Dispatch is not an ELD, not a freight broker, and not a compliance product. Bring your own ELD. We run the dispatch and the driver app.

Frequently asked questions

What is a good driver turnover rate for a small fleet? ATRI's 2026 report puts average 2025 turnover at 32% for fleets with fewer than 26 trucks. If you are well under that, protect it. If you are over it, look at the daily frictions above before you look at the pay sheet.

Does a driver app reduce turnover? We will not quote a number we have not measured. A driver app removes specific frictions drivers name, like chasing photos, lost lumper receipts, and unproven detention. Whether that moves your turnover depends on whether those were your problems.

Is Howdy Dispatch an ELD or a compliance product? No. It is dispatch software plus a driver iOS app. You keep your certified ELD.

Keep the edge you already have

Small fleets keep drivers because the owner and the dispatcher are close enough to fix problems fast. In a tightening market, that edge is worth more than it was a year ago, and it is easy to lose through small daily insults nobody meant.

Run your own lost-driver number this week. Then look at which of those frictions your drivers are dealing with. If you want the paperwork, the photos, and the load status handled on both sides of the load, start a 14-day free trial of Howdy Dispatch.

Driver Turnover Cost for a Small Fleet in a Tightening Market