If a persona is a character, a usage scenario is the plot, a story of how that persona interacts with a proposed information space to accomplish something real. Scenarios describe the future, not the present: they envision an ideal experience unconstrained by existing business or technical limitations, which is precisely what makes them useful for opening up genuine design possibilities rather than just reproducing what already exists.
A well-formed scenario stays focused on behavior, what the persona is trying to do and why, rather than describing specific interface elements like buttons or menus. Writing one generally moves through establishing context, describing the persona's goals and motivations, analyzing the tasks involved, walking through a realistic happy-path version of events, considering what obstacles might arise, exploring more complex variants, and finally defining what a successful outcome actually looks like for that persona.
A scenario that describes which button someone clicks has already quietly become a design specification. A scenario that describes what someone is trying to accomplish is still doing its actual job, leaving the "how" open for design to solve.
Scenario A is well-formed. It establishes context (a notification prompted the visit), motivation (anxiety about cholesterol given recent diet effort), and what success looks like (enough clarity to know whether to worry or relax), all without dictating a single interface element. This leaves the actual "how" open for design to solve, which is exactly what a scenario at this stage should do.
Scenario B has quietly become a specification rather than a scenario, "clicks the blue Results tab," "clicks the dropdown menu," "clicks Compare to Previous" are all interface decisions that should be the outcome of design work informed by a scenario like A, not baked into the scenario itself. Writing scenarios this way forecloses design options before they've been explored, if the team later discovers a better interaction pattern than tabs and dropdowns, a scenario written this specifically actively resists that discovery.Scenario C fails in the opposite direction, "wants to view her health information" is a goal so generic it could describe almost any patient doing almost anything on the portal. It provides no context, no specific motivation, and no way to judge what a successful outcome would actually look like for Maria in this moment. A corrected version of Scenario B's opening sentence might read: "Maria wants to see whether her cholesterol has changed since her last test", behavior and goal, no interface. A corrected version of Scenario C might read: "Maria wants to know if a specific recent lab result is normal or something she should be concerned about", specific enough to be useful, but still free of any interface assumption.