When is Excel still enough for transport operations?
There is no universal truck, job or user count at which Excel suddenly becomes unsuitable. A spreadsheet can remain appropriate when it has a clear owner, one controlled version, input rules and a bounded workflow that the team can review reliably.
A TMS becomes worth evaluating when the same transport order is repeatedly entered into different tools, job status depends on manual synchronization, or billing and decision data consistently arrive too late. These are signals to measure, not proof that software must be purchased immediately. Start with what a TMS is if you need the category boundary first.
The ICAEW principles for good spreadsheet practice take the same risk-based approach: assess how critical the process is and whether a spreadsheet remains the suitable tool before choosing an alternative.
Compare governed Excel with the TMS you can actually deploy
A fair comparison does not put an emailed workbook against an idealized TMS. Compare the Excel configuration your team can govern with the system a supplier can demonstrate for your workflow.
| Criterion | Governed Excel | TMS under evaluation |
|---|---|---|
| Purpose | calculations, analysis, lists and flexible models | repeated operating workflow, within the configured product scope |
| Authoritative record | needs a named owner, structure and rules | depends on the implemented data model and objects |
| Input controls | Data Validation can restrict direct entry | validation must be demonstrated on your cases |
| Collaboration | co-authoring under documented Microsoft 365 conditions | roles, permissions and concurrent work must be tested |
| Change visibility | Version History and bounded Show Changes | audit history varies by product and configuration |
| External data | Power Query can import, transform and refresh supported sources | integrations and direction of transfer must be confirmed |
| Transactional workflow | designed and maintained by the team | may use linked states and records if the product supports them |
| Governance effort | workbook documentation, testing, protection and maintenance | implementation, administration, support and configuration maintenance |
Modern Excel has meaningful controls, but they are conditional. Microsoft co-authoring requires supported Microsoft 365 versions and formats, OneDrive or SharePoint Online storage and AutoSave; an unsupported version can lock the file. Version History has the same cloud-storage boundary.
Data Validation helps with direct entry, but copying, filling, formulas and existing values can bypass or hide warnings. Show Changes is not an immutable audit log: it omits some edit types, can contain gaps and can be cleared. Power Query can connect, transform and refresh data, but it does not turn a workbook into a transactional transport engine.
A TMS does not have one universal feature list either. Even Oracle’s category overview notes that execution features vary and that products may be standalone or integrated with other systems.
Six workflow signals that spreadsheets are becoming the operating system
Do not mark these signals by intuition. Find a recurring case and measure it.
| Signal | Control question | What to measure |
|---|---|---|
| The authoritative record is unclear | Can the team name the current record immediately? | search and confirmation minutes |
| The same data is re-entered | Where are customer, route, rate or vehicle details copied? | entries and minutes per transport order |
| Status needs manual synchronization | How many calls or messages are needed for current status? | interventions and response time |
| Billing or reporting is reconstructed | How many files and messages feed the output? | minutes, corrections and sources used |
| Profitability arrives too late | When can revenue and cost be compared? | delay until usable information |
| Continuity depends on one person | Can another colleague take over without private explanations? | handover time and clarification questions |
For the profitability signal, the transport cost-per-kilometre framework shows which figures must belong to the same period. For vehicle and document records, continue with the fleet record-keeping guide.
Test whether governance can fix the problem first
Before migrating, run a spreadsheet-governance test:
- Name the single workbook and owner. Document its purpose, inputs, outputs, sources and change responsibility.
- Use compatible collaboration. Check Excel versions, file format, cloud storage and AutoSave for every user.
- Control inputs. Use lists, rules and error messages, then test copying, formulas and existing values separately.
- Reduce manual transfer where appropriate. Power Query may help with supported imports and refreshes; do not assume two-way write-back.
- Document and test. ICAEW recommends clear structure, controls, peer review and tests using extreme or impossible inputs.
- Protect access and formulas. Permissions and protection reduce accidental changes but do not replace review.
- Retain history with realistic expectations. Use Version History and treat Show Changes as bounded visibility, not complete audit evidence.
If those controls solve the recurring issue at acceptable effort, keep Excel. If workbook governance becomes a parallel software project while operational state remains fragmented, there is a measurable reason to test dedicated software.
Measure one complete operating and billing cycle
Observe at least one complete operating and billing or reporting cycle. If billing closes monthly, include month-end; mark ordinary, peak and exceptional days separately.
| Log field | What to record |
|---|---|
| Event ID | one stable identifier for the transport order, invoice or incident |
| Participant | person and role performing the task |
| Trigger | what started the task |
| Transfer | source and destination of the data |
| Activity | entry, search, confirmation, correction or reporting |
| Hands-on time | minutes actively worked by that participant |
| Elapsed time | waiting until the result, reported separately |
| Cause and outcome | why the intervention happened and what was corrected or delivered |
Two colleagues working concurrently for 30 minutes use 60 person-minutes, not 30. If one person overlaps two activities, allocate each minute once. Remove duplicate entries using event ID, participant and time window.
Strictly illustrative arithmetic: two roles × 30 minutes per day each × 22 observed days = 22 person-hours. This is not a market benchmark and says nothing about wages, savings or payback.
Compare attributable costs over the same future period
Choose a future evaluation period P made of N complete observed horizons H:
P = N × H
Use parallel pilots or periods that are sufficiently comparable in volume, transport-order mix, roles, service scope, obligations and configuration maturity. When workload differs, normalize only the observations that genuinely vary with workload, using their evidenced driver: transport order, invoice, document or another disclosed unit. Segment materially different job types rather than hiding them in one average.
Do not scale fixed or nonlinear costs as though they were order-variable. Carry each item into P by its real driver: setup once, subscriptions by time, minimum commitments by contract, tiered charges by the applicable tier and labour or correction costs by measured units.
For each configuration X — governed Excel or the selected TMS — calculate:
C_X(P) = I_X + R_X + L_X + Q_X + F_X + G_X
| Symbol | Future attributable cost in P |
|---|---|
| I | setup, migration, training and integration paid once |
| R | recurring software, cloud, integration and support payments |
| L | overtime, outsourcing, temporary staffing or other cash labour actually attributable to the scenario |
| Q | out-of-pocket correction cost, excluding labour already in L |
| F | incremental financing charges and realized cash losses, net of recoveries |
| G | external administration or governance payments not included elsewhere |
Count each cost once. Exclude sunk and common costs from the difference and disclose material common items separately.
Treat VAT recovery as an explicit assumption and confirm it with the company’s accountant for the relevant jurisdiction, tax position and evaluation period. In C_X(P), exclude only the portion confirmed as recoverable and include any non-recoverable portion consistently. The European Commission’s VAT deductions guidance is an official starting point; national rules and the company’s circumstances control the result. This calculation compares only the attributable cost categories in the table, not opportunity value or the timing of receipts and payments. Record VAT payments, offsets and refunds separately in the cash-flow forecast.
The cost difference is:
Δcost = C_TMS(P) − C_Excel(P)
A negative result means the selected TMS scenario has lower attributable cost over P; a positive result means higher cost. Include one-time I once in the main total. For presentation only, E = I ÷ N can show setup cost per H, but E must never be added again.
Calculate C(P) before showing an all-in per-order figure. That figure combines fixed and variable costs and will change with utilization.
Report capacity and timing separately from attributable cost
| Output | Interpretation |
|---|---|
| Cost difference | the model’s attributable cost categories, calculated consistently over the same period |
| Capacity | person-hour difference; not a cost reduction unless payroll or paid external labour changes |
| Opportunity value | separate estimate with a disclosed rate and credible redeployment |
| Service time | elapsed time to status, invoice or report, not hands-on labour |
| Cash flow | timing of payments, receipts and VAT recovery, plus evidenced financing cost; an overdue invoice balance is not itself a loss |
Do not call the result guaranteed savings or ROI before a pilot confirms the workflow, costs and residual work. A problem is addressable only when the evaluated configuration demonstrably handles it; otherwise retain it as residual.
What must a TMS prove in a pilot?
Do not buy a generic feature list. Tie each measured problem to an acceptance test:
- actual users, roles and permissions;
- data import, quality and ownership;
- handoffs between order, status, documents, invoice and reporting;
- required integrations and direction of data exchange;
- export and data portability;
- manual work remaining after configuration;
- ownership of implementation, training and support.
A TMS does not inherently replace accounting or ERP, telematics, compliance advice or Excel used for ad-hoc analysis. Ask the supplier to demonstrate the boundaries and connections between systems.
Keep Excel, pilot a TMS, or split the jobs?
| What the evidence shows | Reasonable decision |
|---|---|
| The workbook is controlled and recurring problems are absent | keep Excel and maintain governance |
| Recurring signals exist but evidence is incomplete | strengthen Excel and run a measured pilot |
| The repeated workflow remains fragmented and the pilot solves the measured cases | move that workflow to the TMS and retain Excel for analysis |
The decision is not “Excel or TMS forever.” It is choosing the right tool for each kind of work, based on an observed process and a demonstrated configuration.
