Referentieklasseforecasting

Referentieklasseforecasting, also known as reference class forecasting, is a prediction technique that improves the accuracy of estimates by comparing a project or decision with a group of similar, past projects. The method was popularized in the 1990s by economist Daniel Kahneman and his colleagues, who identified it as a remedy to the planning fallacy—a tendency for planners to underestimate costs and timelines and overestimate benefits.

The process begins by defining a reference class that captures projects sharing key characteristics such as size, complexity, technology, and industry context. Historical data from this class are then aggregated to determine statistical distributions for outcomes of interest, such as cost overruns, schedule delays, or performance metrics. An individual project is placed within the reference class, and its estimate is adjusted based on where it falls within the distribution. For example, if a project's preliminary cost estimate aligns with the median of the reference class, the final estimate may be adjusted upward because historical data show that projects with similar starting estimates frequently exceed budgets.

Key advantages of reference class forecasting include its grounding in actual empirical data, its resistance to cognitive biases, and its applicability across a range of fields, from infrastructure and software development to public policy and scientific research. Critics argue that the method requires sufficient high-quality historical data and that it may smooth over unique project-specific factors that could be critical for accurate forecasting.

In practice, reference class forecasting is often integrated into decision-making frameworks such as risk management, earned value management, and strategic planning. By leveraging a structured comparison to past experience, it helps stakeholders produce more realistic, reliable, and transparent estimates.