How do you define a scientific milestone when the result is unknowable?

Why

Instead of writing “demonstrate that target X improves disease phenotype”, write “complete two controlled assays of target X by date Y, verify predeclared quality metrics and decide whether the data justify expansion”. State how many independent samples are feasible, which controls make a negative result interpretable, and the threshold for go, revise or stop. The deliverable is a credible evidence package, including negative findings, limitations and the decision log.

If controls fail, the milestone may be incomplete even with an apparently positive observation. If controls pass and the hypothesis fails, the experiment can still be on time and scientifically successful. Budget review points around dependencies such as reagent arrival and ethics approvals.

Sponsor or grant-specific milestone formats may require quantitative thresholds, but they should be justified, not chosen merely to look precise.