Start with registration and stability
Check the FMCSA registered ELD list yourself rather than taking the vendor's word for it, and check the revoked list too. Providers do get removed, and when that happens you have eight days to replace every device in your fleet.
Then ask about the company. How long has this device been on the list? How many carriers run it? Has the vendor ever had a device revoked? A provider that has been shipping the same certified device for years is a materially different risk than one that launched last quarter.
Vendor failure is a compliance event rather than an inconvenience. Eight days to re-equip an entire fleet is not enough time to shop carefully.
Interrogate the contract, not the feature list
This is where fleets get hurt, and it is the part nobody reviews carefully because it arrives after the decision has effectively been made.
- Term length and auto-renewal. Multi-year terms with automatic renewal and a narrow cancellation window are common and worth negotiating out.
- Early termination fees, and whether they apply per device or per account.
- Hardware ownership. Do you own the devices or are they leased? If leased, what happens at the end of term?
- Price escalation clauses. A rate that rises annually by an unspecified amount is a blank cheque.
- What happens when you add or remove trucks mid-term. Seasonal fleets get punished by rigid seat counts.
- Data export on termination: format, completeness, cost, and time limit.
Support is the feature you'll use
Software gets used by the office during business hours. Support gets used by a driver at a scale house on a Sunday night, and that is the interaction that decides how the fleet feels about the vendor.
Ask specifically, and ask for numbers rather than adjectives.
- Are support hours truly 24/7, including the driver-facing line?
- Is support staffed in your operating region and in the languages your drivers speak?
- What is the median time to reach a human by phone? Not the target, the actual median.
- Can drivers get help directly, or must everything route through the office?
- What is the hardware replacement process and the real turnaround for a failed device?
- Is there an implementation contact for the first 90 days, or are you handed a help centre link?
Ask for the median phone wait time. A vendor that can't produce the number is telling you what it is.
Hardware questions worth asking
Hardware is where the total cost of ownership hides, and where a bad choice generates a slow drip of driver frustration.
- What is the annual failure rate, and who pays for replacements, in warranty and after?
- Does it fit your vehicles? J1939, J1708, and OBD-II are not interchangeable, and mixed fleets need the right cable per unit.
- Is it self-installable? A device requiring a shop visit costs a half-day per truck before you have logged a mile.
- How does it behave with no cellular coverage? Does it cache locally and sync, or does the driver lose the record?
- Does the driver app run on their own phone, or does the fleet supply tablets? Both are valid, and the cost difference is significant.
- How is firmware updated, and can an update be deferred if it lands mid-shift?
Data ownership and integration
Your log data, GPS history, and maintenance records are yours in principle. Whether they're yours in practice depends on whether you can get them out in a usable form.
Ask for API documentation before signing rather than after. A vendor with public, readable API docs expects you to integrate. A vendor who offers to "discuss integration options" often has a CSV export and a professional services quote.
- Is there a documented public API, and is it included or priced separately?
- Are there existing integrations with your TMS, payroll, fuel card, and maintenance systems?
- Can you bulk-export historical data yourself, at any time, without a support ticket?
- Who owns the data, per the contract's actual language?
- How long is data retained after cancellation, and is it retrievable during that window?
Run a real pilot
A demo shows you the software working. A pilot shows you the company working. Insist on one, and structure it to stress the things a demo can't.
- Five to ten trucks, at least 30 days, including a full month-end.
- Pick your least tech-comfortable drivers rather than your most. They are the ones the rollout will depend on.
- Run a deliberate support test. Call the driver line at 2 a.m. and time it.
- Cover a route with known dead cellular zones and confirm the data reconciles afterwards.
- Have a driver practise a roadside data transfer with someone playing the officer.
- Simulate a device failure and time the replacement end to end.
The pilot metric that predicts satisfaction best is not feature coverage. It is how long the vendor took to answer the phone.
Switching providers
If you're replacing an incumbent, sequence it deliberately. Export your historical data before giving notice, since leverage disappears the moment you cancel. Confirm your six-month retention obligation is satisfied by data you hold rather than data the old vendor holds.
Then run both systems in parallel on a subset of trucks for a couple of weeks. Parallel running costs a little and prevents the failure mode where you discover a gap after the old system is dark.
