Choosing a TMS for a 10 to 50 truck fleet

Carriers between roughly ten and fifty trucks are in the hardest spot in the software market. Below ten, a spreadsheet genuinely works. Above a hundred, you can afford enterprise software and the person to run it. In between, you have all the complexity and none of the overhead to absorb it.

Here is what actually matters when you shop.

Start from the week, not the feature list

Before looking at any demo, write down your week. Not the idealized version — the real one:

• Monday morning, who decides which driver takes which load, and what do they look at to decide?
• When a driver calls about a check engine light, where does that go?
• On Thursday, who chases missing PODs, and how do they know which are missing?
• On Friday, what does it take to produce settlements, and what makes it take longer?

Then evaluate software against that document. Most demos are organized around the vendor's module list, which tells you what they built, not whether it fits how you work.

The features that actually carry weight

A dispatch board you can read in one glance. Not a list of loads — a timeline showing every driver and where they are in the week. If you have to click into records to understand your own week, dispatch will keep running on a whiteboard and the software becomes a data-entry tax.

Settlements that handle real agreements. Percentage splits are table stakes. Ask specifically: can I charge a driver an exact dollar amount of a repair? Can I spread it over several weeks? Can I mark a lumper as billed to the broker so it never touches driver pay? These are where spreadsheets win and weak software loses.

Documents attached to the thing they belong to. The rate confirmation belongs to the load. The shop invoice belongs to the repair and to the unit. The medical card belongs to the driver. If documents live in a general file store, you are just running Dropbox with extra steps.

Maintenance as a record, not a filing cabinet. A closed repair should tell you the reported symptom, the actual cause, what was done, the mileage, and what it cost. That history is what tells you a truck is becoming a problem before it strands someone.

Compliance with dates that warn you. CDL, medical, permits, registration, inspections. The value isn't storage — it's the alert three weeks out.

The features that get oversold

AI everything. AI document reading is genuinely useful — dropping a rate confirmation and getting structured fields back saves real minutes per load. AI that promises to optimize your dispatching is, for a fleet your size, usually a dashboard with a confident tone. Ask to see it run on your actual documents.

Load boards built in. Nice, but you almost certainly already have DAT or Truckstop and you won't stop using them.

Driver apps. Valuable if your drivers will actually use them. Ask the vendor what percentage of drivers at similar carriers log in weekly. The honest answer is often disappointing, and a driver app nobody opens is worse than a phone call.

Accounting suites. Most carriers keep QuickBooks. A TMS that exports cleanly beats one that tries to replace your accountant's tools.

Questions that reveal the truth quickly

• "Show me a settlement with a split repair charge and a company-paid lumper." If they have to explain a workaround, settlements are weak.
• "What happens when a load has three pickups and two drops?" Multi-stop is common and surprisingly often bolted on.
• "How do I find every repair on truck 104 last year?" Tests whether maintenance history is real.
• "What does it cost to add my safety director?" See the pricing article — this one predicts data quality.
• "Can I get all my data out, in a usable format, today?" Vague answers are a red flag about lock-in.

Migration is the part people underestimate

Whatever you pick, you have to get your fleet into it: trucks, trailers, drivers, customers, and enough history to be useful. Ask what import looks like. CSV import for the core lists should be standard. If the vendor's answer is "our team will handle it" with a four-figure fee, factor that in.

Plan to run parallel for two or three weeks. The week you cut over is not the week to also trust it for payroll.

The honest test

After a two-week trial, ask your dispatcher one question: would you go back to the spreadsheet?

If they hesitate, the software isn't fitting the work. No feature list survives that answer.