Overview
Scientific workflows combine protocols, samples, instruments, observations, and decisions. When software treats them as unrelated forms, context disappears between steps and the record becomes a collection of entries rather than a useful account of the experiment.
The difficulty is not only recording what happened. The system must help a team understand what was planned, what changed, what evidence was produced, and which result can be used in the next decision.
The problem
Laboratory work rarely follows a perfectly linear path. Samples may be split, runs may be repeated, protocols may be revised, and a result may need review before it can be used. A form-based product often makes these relationships implicit.
That forces scientists to carry context in notebooks, spreadsheets, or memory. It also makes implementation brittle because each new exception becomes another disconnected status or workaround.
The approach
The design should model the experiment as a connected sequence before deciding which screens are needed. Protocol versions, sample relationships, instrument runs, observations, deviations, and approvals need clear ownership and visible links.
Normal and repeated paths are mapped together. A failed run is not treated as a dead end; it is connected to the reason for repeating the work and to the record that ultimately supersedes it.
The decision
Software should preserve the structure of the work even when the interface keeps the immediate task simple. Users can focus on the next action while the system maintains relationships, provenance, and history in the background.
Evidence package
The evidence includes workflow maps, domain models, interface prototypes, usability findings, and tests for incomplete, repeated, corrected, and interrupted experimental paths. Reviewers should be able to follow a result back to the protocol and material that produced it.
What the case shows
Laboratory software works when its structure preserves how an experiment actually unfolds. The interface becomes clearer because the underlying model respects the relationships scientists already use to reason about their work.