1.6 Defining success

A project needs more than a completion deadline. Success criteria should address both analytical quality and practical usefulness. Examples include:

• every reported value can be reproduced from documented data and code;
• projected demand is calculated from compatible populations and rates;
• the selected model performs better than a relevant baseline on unseen data;
• each recommendation is supported by client evidence and has a measurable evaluation plan;
• the final audience can understand the central conclusion without reading the technical appendix; and
• confidential information is handled according to all applicable requirements.

A criterion such as “produce a professional report” is too vague by itself. A stronger criterion specifies what professional quality means in the context of the project.

Quality standard: A project question should identify the population or unit, the outcome of interest, the comparison or time frame, and the intended use whenever these are known.