The subscription is the visible cost and frequently a minority of the total. Here is what proposals leave out, roughly in order of how often it causes a budget overrun.
1. Internal effort
The largest omitted cost in most implementations. Someone from your operation must define requirements, cleanse data, make configuration decisions, test, train and support go-live. For a mid-sized implementation this is a substantial fraction of a full-time role for several months, plus workshop and operational staff time.
Budget it explicitly. Projects that treat internal time as free are the ones that slip, because the people concerned still have their day jobs.
Single sign-on is regularly positioned as an enterprise feature at a significant premium, despite being a basic security control.
2. Data preparation
Cleansing the asset register, resolving duplicates, verifying geocodes, reconciling with finance, standardising part numbers. Almost never in the vendor's scope, always on the critical path. See data migration.
3. Integration
Quoted vaguely or excluded entirely. A bespoke ERP integration frequently costs as much as or more than the first year of licences, and it is where schedules slip. Get it scoped and priced as a separate line with defined interfaces.
4. Hardware and its lifecycle
Devices, mounts, chargers, cases, spares — plus replacement every two to three years, installation, de-installation at vehicle disposal, and the spare pool needed to cover failures.
5. Connectivity
SIM cards, data plans, roaming for cross-border operation, and overage charges. Video telematics uses substantially more data than tracking, and overage is a common unpleasant surprise in the first quarter.
6. Annual uplift
An uncapped annual increase compounds. Over a five-year term this can add materially to the total, and it is entirely negotiable at signature and impossible afterwards.
7. Premium modules
Single sign-on, API access, advanced reporting, sandbox environments and additional entities are frequently priced outside the base tier. Confirm which of your requirements are in the quoted tier, in writing.
8. Training
Beyond the initial sessions: new starters, refreshers after releases, and materials. High-turnover driver populations need continuous onboarding, which is an ongoing cost rather than a project one.
9. Support tier
Standard support may mean business hours in a different timezone with a next-business-day response. For an operation that starts at 04:00, that is inadequate. Priority support is usually an additional percentage.
10. Configuration changes
Some vendors charge professional services rates for configuration changes after go-live. If your business changes — new depot, new pricing structure, new vehicle class — that becomes a recurring cost. Establish what you can change yourself.
11. Reporting and analytics
Vendor reporting is often adequate operationally and inadequate analytically. Fleets frequently add a BI tool, a data warehouse and the effort to build and maintain the pipeline.
12. Parallel running
Running old and new systems simultaneously means paying for both, plus the duplicated effort. Budget a defined overlap period and hold to it.
13. Productivity dip
Real and temporary. Expect reduced throughput for the first two to six weeks after go-live. In a delivery operation that is overtime, hire vehicles or service failures — a genuine cost that is never in a proposal.
14. Exit
Data extraction, format conversion, archive hosting, de-installation, and the cost of running two systems during migration to a successor. Ask about extraction cost and format before signing, not when you are leaving.
Surfacing them before signature
Ask each vendor to complete a five-year total cost table:
| Item | Y1 | Y2 | Y3 | Y4 | Y5 |
|---|---|---|---|---|---|
| Subscription | |||||
| Implementation | |||||
| Data migration | |||||
| Integration | |||||
| Hardware | |||||
| Connectivity | |||||
| Training | |||||
| Support tier | |||||
| Estimated uplift | |||||
| Exit / extraction |
Then add your own lines for internal effort, parallel running, productivity dip and reporting infrastructure. The resulting total is frequently a multiple of the headline subscription, and it is the only number worth comparing between vendors.
Questions readers send us
How much should we budget beyond the subscription? Enough that the total is not a surprise: for a complex implementation with integration and migration, first-year non-subscription costs frequently equal or exceed the subscription itself. Build your own model rather than applying a rule of thumb.
Are these costs vendors hiding things deliberately? Usually not — proposals cover what the vendor supplies, and internal effort, productivity dips and reporting infrastructure are genuinely outside their scope. The omission is structural rather than deceptive, which is why the buyer has to add them.
Can we avoid implementation costs by doing it ourselves? Partly, for simple deployments. It converts an external cost into an internal one rather than removing it, and self-implementation without experience frequently produces configuration decisions that are expensive to unwind later.
What is the single most underestimated cost? Internal effort, consistently. The second is data preparation. Both are invisible in procurement and both determine whether the project lands on time.
How do we protect against future cost increases? Negotiate uplift caps tied to a published index, fix your feature tier contractually, agree pricing for adding units mid-term, and confirm exit and extraction terms — all before signature, when you still have leverage.
