Performance targets, tolerances and exceptions

PRINCE2 7 describes seven performance targets that every project must manage: benefits, cost, time, quality, scope, risk and sustainability. Targets are set for each level of management, tolerances define how far performance may deviate before the next level up must be told, and exceptions are raised when a tolerance is forecast to be exceeded. PRIN-110 covered the principle of managing by exception and the basic arithmetic of tolerances. This lesson looks at the same machinery from the people side: who sets and agrees targets, how targets shape behaviour, how to forecast honestly and how exceptions are handled as decisions between people rather than as paperwork.

The seven performance targets

The table below summarizes the seven targets, with typical ways of expressing a target and its tolerance. Sustainability is new in the seventh edition and gets a lesson of its own next.

Not every target needs a numeric tolerance at every level. A work package may have only time and cost tolerances, with quality defined through product descriptions. A project-level benefits tolerance may be expressed as a range. What matters is that each level knows which targets it is managing to and what would trigger escalation.

Who sets and agrees targets

Targets flow down through the management layers: the business layer sets project tolerances, the board sets stage tolerances and the project manager sets work package tolerances. In people terms, though, good targets are not simply imposed. They are discussed. A project manager who accepts an unrealistic stage target without saying so is storing up an exception. A board that sets tolerances without understanding the plan may set them too tight, causing needless escalations, or too loose, losing control. The best practice is a conversation at each stage boundary in which the project manager presents the stage plan, forecasts and risks, and the board sets tolerances it can live with.

How targets shape behaviour

People respond to the targets they are measured against, sometimes in ways nobody intended. If a team is judged only on time, it may meet dates by quietly reducing quality. If a project manager is judged only on cost, they may avoid spending on risk responses that would save money later. An old observation, often called Goodhart's law, warns that when a measure becomes a target it can cease to be a good measure. PRINCE2's use of seven targets is partly a defence against this: by watching several targets together, the board can see when one is being met at the expense of another.

Forecasting honestly

Management by exception depends on forecasts, not just actuals. An exception should be raised when a tolerance is forecast to be exceeded, not when it has already been exceeded. Forecasting therefore needs both a method and honesty. Earned value is one widely used method: comparing the value of work completed with the cost spent and the schedule planned gives indices that can be used to project the final outcome. The method is only as good as the inputs, and inputs are only as good as people's willingness to report real progress.

The same logic applies to time. The schedule performance index, earned value divided by planned value, gives a rough indication of whether the stage will finish late. Here SPI is 270,000 ÷ 300,000 = 0.90, suggesting the work is running about 10 % behind plan. Earned value indices are a guide, not a certainty, and the project manager should combine them with the team managers' judgment about remaining work.

The escalation path

When a tolerance is forecast to be breached, the escalation path is the same at every level: the person who holds the tolerance raises the matter with the level that set it. A team manager raises an issue with the project manager. A project manager sends an exception report to the board. The board, if project tolerance is threatened, refers the matter to the business layer. At each step, the person escalating should explain the cause, the impact on all targets, the options and a recommendation.

The people dimension is timing and tone. An exception raised early, with clear options, is a sign of good management. One raised late, or presented defensively, erodes trust. Boards should respond promptly; an exception report that waits three weeks for a decision may itself cause further breaches.

Responding to an exception

When the board receives an exception report, it has several options. It can ask for an exception plan to replace the current stage plan, approve a change in tolerances, accept a reduction in scope, ask for more information, or, in serious cases, decide that the project is no longer justified and request premature closure. The decision should be based on the business case: which option gives the best remaining value?

This example shows that exceptions are decisions involving judgment, risk and trust. The numbers inform the decision, but the board must also weigh confidence in the forecasts, the team's capacity and the effect on stakeholders.

Targets across organizations

When suppliers are involved, targets may also appear in contracts, and the two do not always align. A supplier's contract may fix a price and date that become the supplier's targets, while the customer's stage tolerances allow different flexibility. Mismatches cause friction: the customer board may want to trade time for quality while the supplier is contractually bound to a date. Aligning work package tolerances with contract terms where possible, and agreeing in advance how contract changes will be handled, reduces this friction.

Benefits and risk tolerances in practice

Benefits and risk tolerances are the hardest to manage day to day because they are less tangible than cost and time. Benefits tolerances are usually expressed as a range for the expected benefits in the business case, and the forecast is reviewed at each stage boundary. If the expected benefits fall below the tolerance, for example because a key user group will not adopt a product or because market conditions have changed, the executive must consider whether the project is still justified. This is a people decision as much as a financial one, because it often means telling sponsors and teams that the value they hoped for has shrunk.

Risk tolerance describes how much aggregated threat exposure the organization will accept, often expressed as an expected loss or as a limit on particular kinds of risk, such as no risks of serious harm to people above a low likelihood. Risk appetite comes from the business layer and reflects the organization's culture: some organizations accept considerable risk in pursuit of innovation, others very little. When a project's risk profile moves beyond the agreed tolerance, the board must decide whether to accept the higher exposure, invest in responses or change the project.