The IT business case must remain alive after launch

The proposal looked profitable until someone asked where the hours went

In this illustrative 2021 scenario, Antoine proposes an invoice-processing project. The supplier predicts 2,000 hours saved annually. At €60 per hour, the proposal claims €120,000 of benefit. Sarah, the management accountant, asks which spending will fall, which capacity will be reassigned and who can demonstrate it. The silence does not invalidate the project; it reveals different categories of value being added without explaining how they will materialise.

Written in 2026, this retrospective examines 2021 investment decisions. Figures, people and assumptions are teaching examples, not a CYTIZEN engagement or promised return. A decision-ready business case compares alternatives, separates cash and capacity, checks remaining costs and allows revision when actual work disproves an assumption.

Define the no-project option before selling the project

Sarah requests annual volume, invoice categories, manual steps and exceptions. Antoine observes straightforward invoices, non-purchase-order invoices and recent corrections. Processing effort includes chasing, checking and rework, not just screen entry. Difficult cases remain in the sample. Separate working time from waiting: removing three days of delay does not release three days of salary, although it may improve an important business outcome.

No project does not mean a business frozen for five years. Include expiring contracts, volume growth and improvements already agreed. Do not unfairly load unavoidable costs onto the project or credit it with independent savings. Use consistent populations, horizons and calculation conventions across options. Unmeasured assumptions retain an identified owner and confirmation requirement.

A completed case separates categories of value

In practice: the benefit lost in checking

The pilot reduces entry effort but adds checking for invoices without purchase orders. Antoine decomposes the 2,000 hours into automated work and tasks that remain. Sarah asks whether recovered time is concentrated in a team able to use it or scattered in minutes across many people. Fragmented time cannot automatically support headcount reduction. A manager can instead commit it to reducing backlog, with observable results.

The proposal therefore changes from a self-funding savings claim to an investment in quality and processing capacity. That may be legitimate, but requires its own criteria. Finance asks how much additional work will be completed, which defects will decline and which spending remains. The committee may accept this value, narrow scope to purchase-order invoices or select cheaper procedural improvements. Returning to the decision is not a failed pilot.

Make the timing calculation consistent

Investment precedes benefit. Year one can include ramp-up, parallel operation and training. If only half the central €100,000 annual benefit appears while €60,000 operating cost is already incurred, first-year economic effect is negative €10,000. Combined with €180,000 initial investment, the cumulative position is negative €190,000 at the end of year one. Subsequent years add €40,000 if central assumptions materialise.

This illustrative undiscounted model reaches economic break-even after 5.75 years, not the 4.5 years obtained by ignoring ramp-up. Cash does not break even with €40,000 avoided spending against €60,000 recurring cost. State which measure is being calculated. For a real decision, Finance confirms discounting, tax and financial conventions; the supplier must not invent them to improve results.

Test assumptions capable of changing the decision

Sarah varies reusable effort, checking workload and reducibility of external services. The cautious scenario assigns no value to capacity and retains only contractual savings: negative €20,000 annually before investment. A favourable scenario reassigns all 2,000 hours to measured work, giving €160,000 gross economic benefit and €100,000 after operating costs. This remains a favourable scenario rather than silently replacing the central assumption.

Sensitivity analysis matters when it changes authorisation. If acceptance requires more than 50% reusable capacity, the pilot must measure it and management confirm reassignment. If checking consumes the saving, redesign exception handling before expansion. If the supplier contract cannot be reduced, remove the cash saving. Multiple scenarios that affect no decision simply present the same optimism more elaborately.

Give every benefit an owner

Procurement owns contractual reduction; the accounting manager owns productive reuse of capacity; process quality and Finance validate control outcomes. Antoine coordinates but cannot substitute for them. Each benefit records source, baseline, observation date, population and enabling action. Time measures retain sampling method and exception frequency. Cash is reconciled against amended contract, supplier invoice and observed spending.

The case also records extra work, new incidents, supplier dependency and exit cost. Speed may reduce control; quality may justify higher cost. Do not force every dimension into an artificial monetary figure. The committee sees what it is buying, what it is not buying and which disproved assumption requires renewed approval.

What benefit means in different sectors

Manufacturing and pharma may prioritise complete records, Quality capacity and exception handling over entry speed. Include required validation and change work in cost. Traceability improvement needs evidence across the flow, not an inflated hypothetical penalty used to manufacture profitability.

Banking and insurance separate handling cost, case quality and customer outcome. Shorter delays do not automatically create additional revenue. Controls and post-incident work remain in the model. Services and public administration may use released capacity to improve access or reduce backlog. Owners specify populations, case types and expected outcome. Service value remains legitimate even without a lower budget.

Keep the business case alive after approval

When volume, scope or cost changes materially, Sarah updates affected assumptions and shows the difference from approval. The committee compares remaining future cost and benefit without being trapped by sunk investment. The original case is neither erased nor presented as still true. A sound decision explains why it was taken, what has since been learned and which alternative is now preferable. Managing the business case means preserving that ability to choose.

Sources and method

Primary sources checked on 4 October 2026. The year identifies the period being examined; this retrospective was written in 2026. Later documents provide present-day comparisons, not knowledge attributed to that period. People, scenarios and numerical examples are illustrative teaching material, not results of a CYTIZEN engagement.

• HM Treasury — Green Book, editions and framework checked in 2026
• Government Project Delivery — Quick guide to benefits management, July 2026