The whole-life cost of a factory system beyond the licence line

The visible fraction and what sits beneath it

Beneath the software price sit implementation services, integration development, data cleansing and loading, environments for testing and training, hardware on the floor, network coverage into areas that never needed it, training delivery, and the travel and time of everyone involved. Vendors quote the parts they supply. The categories they do not supply are the ones that vary most between sites, and they are where comparisons between competing proposals usually become misleading, because each quotation draws the boundary in a different place.

Internal time is real money and rarely counted

The planner spending half their week in workshops, the supervisor testing transactions, the storeman recounting inventory before cutover, the engineer rewriting routings — none of that appears in a proposal and all of it is either paid overtime, deferred work, or output the plant does not produce. Costing it does not make it more expensive; it makes the case honest and it exposes whether the organisation can actually supply the effort. Projects that skip this exercise discover the constraint later as slipped dates, which cost more than the backfill would have.

The costs that start on the day you go live

Recurring spend includes support and maintenance, version upgrades and the regression testing they require, interface maintenance whenever either end changes, configuration work as products and processes evolve, and retraining as people leave. Interfaces are the item most often forgotten: every connected system has its own release cycle, and each one can break a link you did not touch. Budgeting a standing allowance for integration work rather than treating each break as an unplanned incident is more accurate and much less disruptive to argue about.

Hardware in an environment that destroys hardware

Terminals, scanners, printers, tablets and access points on a shop floor face dust, moisture, vibration, temperature and being dropped. Consumer-grade equipment is cheap and fails; ruggedised equipment costs more and lasts. Beyond the units themselves, budget for spares held on site so a failed scanner does not stop a line, consumables for label printers, mounting and enclosures, cabling into areas with none, and wireless coverage inside a metal-clad building, which frequently costs more than anyone expected during the survey.

Costing the workarounds you agreed to keep

Most go-lives conclude with a list of things the system does not do, each handled by a spreadsheet, a duplicate entry or a weekly reconciliation somebody performs by hand. Individually each is minor. Collectively they are a permanent staffing cost that never appears in any comparison, and they usually grow. When evaluating options, ask what would remain manual under each, cost that effort as recurring, and revisit the list after go-live to see which items could now be closed. Nobody schedules that review, which is why the workarounds outlive the project team.

Frequently asked questions

How should we compare a subscription against a perpetual licence?

Model both over the same period, and include what each assumes about the other categories. Perpetual licences carry annual maintenance, infrastructure, and upgrade projects that are separate purchases. Subscriptions bundle some of that and expose you to price changes at renewal. Ask what happens if you stop paying: perpetual leaves you with unsupported software you may still run, subscription leaves you with an extraction deadline. That difference matters more to a manufacturer than the arithmetic.

What gets underestimated most often?

Data work and internal time, in that order. Cleansing an item master and rebuilding routings takes far longer than any plan allows because the effort is not proportional to record count but to how many judgements a person has to make. After that come integration, shop floor hardware and network coverage, and the testing burden of the first upgrade. Underestimating these does not change the eventual spend; it changes whether the project is judged a success.

Should we budget for a second phase from the start?

Yes, and name it explicitly, because it is going to happen whether it is budgeted or not. First implementations concentrate on getting transactions working, and the capability that justified the investment — scheduling, analysis, quality integration — is usually deferred. A defined follow-on with its own funding and a review of what people are still doing manually converts a deferred benefit into a delivered one. Without it, the deferred items become the permanent state.