Project management: when should you test before freezing requirements?
A project must turn expectations into commitments without treating fragile assumptions as established facts. The management question is specific: which requirements should be tested before they are frozen in a specification?
What the literature establishes
Kathleen M. Eisenhardt and Behnam N. Tabrizi (1995, Administrative Science Quarterly) distinguish, in computer product development, between compressing known stages and an experiential approach involving iteration and testing. Their findings indicate that the value of these approaches depends on contextual uncertainty. Accelerating a project therefore does not always mean executing a predefined sequence faster: it may require learning earlier what needs to be developed.
The mechanism: making an assumption testable
Stefan H. Thomke (1998, Management Science) analyses experimentation as a design activity whose organisation and tools influence the cost and time required to learn. A prototype is not merely an incomplete version of the deliverable: it can help distinguish between alternatives before a more costly commitment. For the project manager, the task is to specify the uncertainty each test should reduce, rather than simply increasing the number of demonstrations.
The premature conclusion
One might conclude that prototypes should replace specifications. These studies do not justify that opposition: experimentation helps establish some requirements, while documentation remains necessary to coordinate delivery and decide on acceptance. A regulatory obligation does not become negotiable because users like a prototype. (our executive and employee training programmes)
What these data cannot resolve
These articles concern product development and do not provide direct validation for every service project. A trial conducted in a simplified environment may leave integration, security or large-scale operational constraints undetected. Common practice adds another difficulty: an approved demonstration is not necessarily a successful test if no acceptance criterion was set beforehand.
A practical check in Genève
Within SHR's Project Management programme, an exercise applicable in Genève is to select a requirement that remains uncertain and record the assumption, test and acceptance criterion before approving it. Depending on the sector, the test could concern a private banking onboarding journey, an approval workflow in an international organisation, a traceability interface in luxury watchmaking or a document handover in commodity trading, using fictitious data where necessary. During a pilot phase, measure the proportion of requirements tested in this way that are subsequently reopened because they do not meet usage needs, retaining the test evidence and reason for reopening in each case. Compare this rate with that of untested requirements over the same period, noting differences in complexity: the comparison informs a local decision without, on its own, establishing causality. To go further: explore the Project Management training in Geneva, or browse our executive and employee training programmes in Switzerland.
In pictures: Project Management in Geneva
- project management
- Geneva
- research
- training geneva