Measurement tells leaders what changed; attribution asks whether the investment materially contributed to the change.
A transformation reports that customer satisfaction has risen. A digital program claims the benefit. At the same time, a competitor exited the market, service volumes fell and another operational initiative changed frontline staffing.
The KPI improved. The benefit may be genuine. The difficult question is whether the program can credibly claim to have created it.
This distinction matters because organisations allocate future capital using the evidence produced by past investments. If programs overclaim benefits, the portfolio learns the wrong lessons. Weak attribution therefore does more than distort reporting. It damages future investment decisions.
The Strategic Context
The 2017 New Zealand Treasury benefits-management guidance explicitly separates benefit measurement from attribution. It recommends testing whether the target would have been achieved without the investment and whether external reasons may have produced the result. It also warns against claiming the same benefit across multiple projects or programs, commonly described as double counting.
The guidance further notes that benefit dependencies may differ from project-delivery dependencies. A benefit can rely on an external activity, another program or business-as-usual action even after the project output itself is complete.
These ideas raise the standard of benefits governance. A benefits register should not become a collection of positive KPI movements. It should contain a defensible explanation of contribution.
What Leaders Commonly Misread
The first misread is correlation equals contribution. A metric moved after implementation, so the program receives credit. Timing alone is not evidence of causation.
The second is one metric equals one benefit. A complex benefit may need several measures, while one measure may be influenced by several initiatives. The relationship needs to be understood rather than assumed.
The third is project scope defines benefit scope. A project can deliver its component successfully while the final benefit depends on operational or external factors beyond the project boundary.
The fourth is every project can claim the strategic outcome it supports. If several initiatives all claim the same enterprise saving, revenue increase or productivity improvement, the portfolio can manufacture value on paper that the organisation receives only once.
The fifth is precision equals truth. A benefit forecast expressed to two decimal places can still rest on weak baseline data, optimistic attribution or untested assumptions.
Reframing the Issue
Benefit evidence has at least four layers:
Measurement: Did the indicator change?
Attribution: How much of the change is reasonably connected to the investment?
Dependency: What else had to happen for the benefit to emerge?
Integrity: Has the same value been claimed elsewhere?
A credible benefit claim needs all four.
This does not mean every program requires experimental proof. In many enterprise settings, controlled trials are impractical or unethical. Attribution is often a judgement supported by evidence rather than a mathematically certain causal estimate. The governance requirement is transparency about what is known, assumed and shared.
The purpose is not to turn benefit reporting into academic causal analysis. It is to stop management from giving an investment more credit than the evidence can support. A proportionate approach can still distinguish a strongly evidenced contribution from a broad outcome that many internal and external forces shaped.
Strategic Analysis: Build a Benefit Evidence Chain
Start with a usable baseline
A benefit target without a reliable baseline cannot show change well. Baseline definition should therefore happen before or early in implementation whenever possible.
If the expected benefit is lower order-processing time, leadership needs a clear definition of current processing time, the population measured, exclusions, seasonality and data quality. Without that, post-implementation improvement can be debated indefinitely.
Define the causal drivers
The NZ Treasury guidance describes value or driver modelling as one way to separate a high-level target into contributing factors. If the benefit is reduced operating cost, drivers might include labour effort, overhead, scrap, energy or external service cost.
This is powerful because projects can then be linked to the specific drivers they are expected to change instead of each claiming the total enterprise outcome.
A hypothetical manufacturing example makes the distinction clear. An ERP program may reduce planning effort, an automation program may reduce direct labour, and a supplier initiative may reduce purchased cost. All contribute to operating-cost reduction, but they should not each claim the same total saving.
Separate contribution from coincidence
Ask two simple questions adapted from the NZ guidance:
- Would the observed result probably have happened without this investment?
- What other factors could plausibly have produced the same result?
The answers should influence confidence in the benefit claim.
If sales increased during a broad market boom, the program may have contributed but cannot automatically claim the whole increase. If throughput improved only on the lines that adopted a new process while comparable lines did not, the attribution argument is stronger.
Make dependencies explicit
Benefit dependencies may sit outside the program boundary. A new analytics platform may depend on managers actually using the recommendations. A new transport service may depend on complementary infrastructure. A workforce system may depend on updated policies and manager behaviour.
Delivery dependencies tell the program what must happen to complete work. Benefit dependencies tell the organisation what must happen to create value.
Related article: Integration Is the Real Work of Program Management
Prevent double counting at portfolio level
Double counting is fundamentally a governance problem. Project-level business cases can look internally consistent while the enterprise portfolio claims the same benefit multiple times.
A central benefits view should therefore reconcile major benefits across investments. Common enterprise targets such as headcount reduction, cost savings, revenue growth, risk reduction and customer improvement require particular scrutiny because multiple programs often contribute to them.
Record confidence, not only target value
An organisation can improve decision quality by recording a confidence assessment alongside the benefit target. This need not be a false quantitative probability. It can distinguish evidence such as:
- directly measured and strongly attributable;
- measured with shared contribution;
- estimated using validated drivers;
- plausible but materially assumption-dependent.
This prevents a forecast from appearing more certain than the evidence allows.
Decision Framework
For each material benefit, use the Benefit Evidence Chain:
| Element | Question |
|---|---|
| Baseline | What was happening before the intervention? |
| Target | What measurable change represents value? |
| Driver | What mechanism is expected to produce the change? |
| Dependency | What else must happen inside or outside the program? |
| Attribution | How much of the result can reasonably be linked to this investment? |
| Reconciliation | Is any other initiative claiming the same value? |
| Confidence | How strong is the evidence behind the claim? |
Governance should challenge the chain, not only the final number.
From Strategy to Execution
Immediate action: select the ten largest benefits claimed across the portfolio and test for baseline quality, overlapping ownership and double counting. This often reveals more value leakage than another round of project-status reporting.
Medium-term capability: create driver models for strategic outcomes shared by several programs. Allocate contribution to component initiatives and make cross-program dependencies explicit. Establish a central reconciliation point for major financial and operational benefits.
Long-term strategic positioning: use realised-benefit evidence to improve future forecasting. When programs consistently overstate attribution or understate dependencies, revise estimation practices rather than treating each miss as an isolated delivery problem.
Related article: Portfolio Reporting Should Change Decisions, Not Produce More Data
Signals to Monitor
Warning signs include identical benefit language appearing in multiple business cases; large savings with no clear baseline; benefits measured only after delivery; programs claiming whole-enterprise KPI movements; unexplained improvements credited entirely to the latest initiative; metrics that move but operational drivers do not; and benefit values that remain unchanged despite material scope reduction.
Positive signals include shared driver models, explicit contribution assumptions, reconciled portfolio benefits, clear baseline ownership and governance willing to report uncertainty rather than manufacture precision.
Questions for the Leadership Team
- What evidence shows this program contributed to the observed change rather than merely preceding it?
- Would the result have occurred without the investment?
- Which other initiatives or external conditions influenced the same KPI?
- Are we claiming the same cost saving, revenue increase or risk reduction more than once?
- Which benefit dependencies sit outside the program manager's authority?
- How confident are we in each major benefit claim, and why?
Closing Perspective
Benefits reporting becomes decision intelligence only when the organisation can distinguish observed improvement from justified contribution. A rising KPI is useful evidence, but it is not automatically proof of program value. The leadership standard should be higher: understand the baseline, expose the drivers, recognise dependencies, reconcile overlapping claims and state confidence honestly. Otherwise the organisation can celebrate benefits that exist only in its reporting system.
About EraNorth Insights
EraNorth Insights publishes practical analysis on strategy, projects, operations, transformation and decision intelligence for professional and organisational use. About EraNorth.