Two projects claim the same saving: reconcile the combined benefit before approving both

Two project teams submit credible business cases. One automates a task and another simplifies the workflow containing that task. Each compares its proposal with today's process and estimates a saving. Adding those standalone estimates can count some of the same work twice, because the second project will operate on a process already changed by the first.

Identify the common operating population before combining the cases. It could be employee hours, transactions, customers, equipment downtime or purchased services. Ask what changes under Project A alone, Project B alone and both together. The combined case needs its own supported outcome; it is not automatically the sum of the two standalone cases.

The UK Government Digital Service's Digital and Data Benefits Framework warns against counting benefits in more than one part of a program and recommends coordinated ownership of the benefits case. That is public-sector guidance, not a reporting requirement for every business. Its underlying question is directly useful here: which measurable outcome belongs in the total once?

Agree the current process boundary, volume, quality requirements and measurement period. A team measuring entry time for completed invoices cannot be combined without explanation with a team measuring all invoice handling, including exceptions and corrections. Map where the activities overlap and which outputs each project changes.

Distinguish removed work from work transferred elsewhere. A workflow change may reduce a preparer's time while adding review or maintenance effort. An automation project may remove manual entry but leave exception handling unchanged. The combined case should include the remaining work and any new tasks, with the same service and accuracy requirements as the baseline.

Retain the evidence behind the baseline. Time observations, transaction records and resource-owner explanations can support the calculation. A broad salary percentage is less informative when the projects affect a small task performed by several roles. The reviewer should be able to trace the claimed improvement to the actual process that will change.

Consider an illustrative process requiring 80 employee hours per month. Project A alone would reduce it to 30 hours, a 50-hour reduction. Project B alone would reduce it to 60 hours, a 20-hour reduction. Adding the standalone claims gives 70 hours, but assume the reviewed combined design requires 25 hours after both projects.

The combined reduction is therefore 55 hours, not 70. If Project A is implemented first, adding B reduces the remaining work from 30 to 25 hours, so B adds five hours of benefit in that sequence. If B is implemented first, A reduces work from 60 to 25 hours, adding 35. The order changes the incremental attribution, while the modeled final outcome stays at 25 hours.

At an illustrative labor value of $40 an hour, 55 hours represent $2,200 of monthly capacity value. The sum of the two standalone claims would show $2,800, overstating that modeled value by $600. Neither figure is automatically a cash saving. A cash claim requires evidence of changed payments or receipts; useful redeployment should be valued through its own supported result.

Some projects do not duplicate a benefit; one makes another possible. A data-cleanup project may be needed before an automation can function reliably. In that case, removing the cleanup cost while retaining the automation's full benefit creates an infeasible case. Record the enabling relationship and evaluate the relevant combination of costs and outcomes.

A shared enabling project may support several later initiatives. Decide how its cost will be represented so the portfolio includes it once, while individual project views remain understandable. The same discipline applies to a common benefit. An attribution rule can support management accountability without changing the total economic outcome.

Ask whether the projects are alternatives, complements or partially overlapping improvements. Alternatives need a comparison of feasible choices. Complementary projects may deliver an additional joint effect that neither achieves alone. Partial overlap requires a combined process estimate such as the 25-hour result above. Naming the relationship helps prevent an arbitrary percentage haircut from standing in for analysis.

Overlap is not confined to employee time. Two customer initiatives may each claim the retention of the same accounts. A control improvement and an insurance change may both claim the full value of avoiding one loss. A procurement project and an inventory project may each describe the same reduction in stock investment. Map the actual outcome and the counterfactual before adding the amounts.

Keep different financial effects separate even when they arise from one process change. Releasing a stock balance is a different event from reducing ongoing storage spending. Faster collection can release cash without increasing revenue. ACCA's relevant-cost approach supports focusing on the future cash flows that change, while preserving their timing and avoiding another label for the same benefit.

For uncertain outcomes, use coherent combined assumptions. Do not assume each project prevents a loss in isolation and then subtract the same loss twice from a portfolio scenario. Ask the risk or operating owner to explain how the interventions interact and what evidence supports the resulting probability or severity assumptions.

Present the cost and outcome of A alone, B alone and the combined option when those are realistic choices. Include implementation costs, ongoing costs and the dates benefits begin. A combined project can still be attractive even after removing duplication, but the decision should use the corrected amount.

Do not cancel a project solely because it receives less credit under one implementation sequence. The five-hour incremental contribution of B after A in the example differs from its standalone 20-hour contribution. The relevant decision depends on which other projects are already committed and what alternatives remain. Explain that decision context instead of treating attribution as an intrinsic score for the project team.

Retain a controlled version of the combined model and its assumptions. Government Actuary's Department guidance on model quality emphasizes appropriate review around model use. For this decision, have a reviewer trace each material benefit from the process evidence into the combined cash or capacity calculation and inspect changes when the program's scope shifts.

Assign one owner to reconcile the relevant operating result across the projects. Individual teams can report completed work and contributions to the change, but the portfolio should not publish separate totals that silently add to more than the observed improvement. Keep the agreed baseline available and explain changes in volume, quality or external conditions.

If only one project is completed, report that state and revise the remaining case. If the final process requires 35 hours rather than the modeled 25, investigate the difference without allowing both teams to retain their full original claims. The financial review should show the result the business obtained, the cost of obtaining it and the specific remaining change still worth making.