Actual action depends on immediate situational adjustment
Aliases: situated adjustment · on-the-spot decision · contextual cue · moment-by-moment decision
What it is
The order of action is continuously organized by on-the-spot cues: who is nearby, whether equipment is available, whether data has just updated, whether time is tight, what trace the previous step left behind. A user acts and watches as they go, treating the current situation as input to the decision, rather than completing a full plan first and then executing it. This card is the positive statement of "a plan is a resource, not a determinant of behavior": that card said a plan cannot determine behavior; this one directly answers what actually does — the continuously shifting situation itself, not some fixed scheme worked out in advance.
Why it happens
A situation can continuously organize action because it constantly offers an actor three things: affordance (what can be done right now), constraint (what cannot be done right now), and opportunity (an unexpected option worth seizing that just appeared). A bed that just became free, a pop-up error, the number of people waiting in line, a colleague's gesture, a message that just arrived — each of these cues changes what to do next in real time, and that change is not a deviation from or a failure of the original plan; it is the actor genuinely treating perceivable current information as decision input. Feedback from the tool itself also keeps participating in this process: the result of one action closes off some previously viable path while opening others that were never considered before. To accommodate this ongoing adjustment, real work is often naturally broken into small, interruptible chunks, so that a suddenly more urgent matter can be inserted at any point and the original task resumed afterward. This way of acting and watching as one goes, always ready to insert and adjust, looks like a lack of planning, but is in fact an efficient strategy for handling a complex, changing situation, not chaos itself.
Where it stops holding
For situated adjustment to genuinely work, two preconditions both have to hold: the key cues in the situation must actually be perceivable, and recovery from an interruption must be low-cost. If the interface's own state is invisible — a user cannot tell whether an object is currently held by someone else, or whether a piece of data has just been updated — situated adjustment has no information to act on, and the actor is left guessing, at which point "on-the-spot judgment" turns into a source of error instead. If an operation, once started, cannot be interrupted, or if the trace of an adjustment made after an interruption is never recorded at all, every adjustment a user makes to cope with a changing situation risks losing progress or creating a conflict. Also, a highly real-time collaborative environment cannot let everyone freely reorder actions purely on their own local judgment — several people adjusting at once, each based on the partial situation they happen to see, can conflict with each other, and such settings still need clear role division and priority rules to coordinate; situated adjustment's freedom should operate within the bounds those rules set, not entirely unconstrained.
Applying it
- Explicitly display the on-the-spot state that affects the next decision: whether an object is currently held, when data was last updated, who the current owner is, and what is still unfinished.
- Support pausing, saving, inserting an urgent matter ahead of the queue, temporarily reordering, and returning to the original task afterward, and log the reason for every adjustment rather than letting adjustments happen without a trace.
- Design an explicit resume point for an interrupted task, so a user who handles an inserted urgent matter can return precisely to the original task's context, without having to reconstruct from memory where they left off.
- How to check: observe real work settings for the whiteboards, verbal reminders, and external spreadsheets users actually rely on — these tools operating outside the system point exactly to which on-the-spot cues have not yet been digitized and need to be explicitly designed into the interface.
Related
- Same group: B4.08.1 Plans are resources rather than determinants of behavior · B4.08.3 The system cannot assume users follow a preset flow
- Nearby: I1 State, Time, and Response · V1 Collaboration and Social Interaction
- Search terms:
situated adjustment·contextual cues·interruption recovery·affordance in context