Finishing Actions
mental-model established
Source: Food and Cooking
Categories: organizational-behaviorsystems-thinking
From: Culinary Mise en Place
Transfers
“Finishing Actions” is the fifth of Dan Charnas’s ten Work Clean principles, derived from the professional kitchen’s discipline of completing tasks through their full cycle before moving to the next one. In the kitchen, this means: when you chop the onions, sweep the trimmings into the bin, wipe the board, and put the prep container in its designated position before pulling out the next ingredient. When you plate a dish, deliver it to the pass, wipe down, and reset your station before reading the next ticket. Every action has a beginning, a middle, and a finishing action — and the finishing action is the one most often skipped under pressure.
Key structural parallels:
-
Incomplete work has carrying costs — the core insight is that a task left at 90% completion does not cost 10% to finish later; it costs more, because returning to it requires reconstructing context. Where did I leave off? What was the next step? Where did I put the half-chopped herbs? In the kitchen, the carrying cost is physical: a half-prepped station clutters the workspace and creates confusion during service. In knowledge work, the carrying cost is cognitive: each unfinished task occupies a slot in working memory (what psychologists call the Zeigarnik effect), reducing capacity for the current task. Finishing actions eliminate these carrying costs by driving each task to zero residual obligation before starting the next.
-
The last 10% is the most important 10% — finishing actions are the cleanup, filing, communication, and handoff that happen after the “real work” is done. In the kitchen: you cooked the dish, but did you wipe the rim of the plate? Did you reset your station? Did you call “ready” to the expediter? In software: you wrote the code, but did you update the documentation? Close the ticket? Notify the stakeholder? The principle recognizes that the final steps of a task are the ones most easily deferred and most costly when deferred, because they are the interface between your work and the next person’s work.
-
Batching is valid only if the batch completes — the principle does not oppose batching similar tasks. A cook who needs to dice onions, carrots, and celery for three different recipes should dice all three consecutively (same tool, same motion, same station setup). But the batch must finish: all three diced, all three stored, all trimmings cleared, board wiped. An interrupted batch — two diced, one still whole, board cluttered — is worse than no batching at all, because the in-progress state of the batch adds its own carrying cost on top of the unfinished items.
-
The Zeigarnik overhead is real and measurable — Bluma Zeigarnik demonstrated in 1927 that incomplete tasks occupy cognitive resources disproportionate to their size. The kitchen quantifies this: a cook with three unfinished tasks is measurably slower on the current task than a cook with zero, even when the current task is identical. The mental model makes this overhead a first-class design constraint rather than a personality trait (“some people are just better at finishing things”).
Limits
-
Not all tasks have clear endpoints — the principle works in the kitchen because culinary tasks have unambiguous completion criteria: the onion is diced or it is not, the plate is wiped or it is not. Many knowledge-work tasks lack this property. When is a design “finished”? When is a document “done”? When is code “clean enough”? Applying finishing-actions discipline to tasks without clear endpoints either forces premature closure (shipping before the work is ready) or enables perfectionist loops (endlessly finishing what cannot be finished).
-
Serial completion is not always optimal — the principle assumes that the best strategy is to finish task A completely before starting task B. But some work benefits from interleaving: writing a first draft, then switching to a different problem while the draft incubates, then returning to revise with fresh perspective. Scientific research often requires running multiple experiments in parallel, not because of poor discipline but because waiting for results creates idle time that parallel work can fill. The principle is designed for the kitchen’s fast-cycle, sequential work and transfers poorly to domains with long feedback loops.
-
WIP limits are not the same as finishing actions — the principle is sometimes conflated with work-in-progress limits from lean and Kanban. Both reduce concurrent work, but the mechanisms differ. WIP limits constrain how much work enters the system. Finishing actions constrain how work exits. A team can have strict WIP limits and still defer finishing actions within each work item (writing code but not tests, building features but not documentation). The principle addresses a finer grain than WIP limits and should not be treated as redundant with them.
-
Urgency can legitimately override completion — in the kitchen itself, finishing actions are sometimes correctly abandoned. When the chef calls an on-the-fly order, the cook drops their current task mid-finish to address the emergency. The principle does not claim that finishing is always correct, but practitioners sometimes apply it rigidly, resisting legitimate interruptions because “I need to finish this first.” The model should include an override mechanism for genuine emergencies, which the kitchen provides but the productivity formulation often omits.
Expressions
- “Finish the action” — kitchen instruction to complete through cleanup, not just through cooking
- “Don’t leave it hanging” — exhortation to complete the finishing actions rather than walking away at 90%
- “Close the loop” — general organizational term for completing the communication and handoff that constitute finishing actions
- “That task is done done” — software team vernacular distinguishing code-complete from truly finished (tested, documented, deployed)
- “Reset your station” — the paradigmatic finishing action in the kitchen, ensuring readiness for the next task
Origin Story
Charnas formulated “Finishing Actions” as the fifth Work Clean principle by observing the discipline professional chefs enforce on their cooks: every task ends with a reset. The station is wiped. The tools are returned. The next task begins from a clean state. Charnas connected this to Bluma Zeigarnik’s 1927 research on incomplete tasks and to David Allen’s Getting Things Done methodology, which similarly emphasizes driving tasks to completion or to a clearly defined “next action” to free cognitive resources. The culinary formulation adds a physical, embodied dimension that Allen’s information-processing model lacks: the finishing action is not just a cognitive commitment but a bodily practice — the hand wiping the board, the arm returning the container to its place.
References
- Charnas, D. Work Clean: The Life-Changing Power of Mise-en-Place (2016) — the principle as formulated
- Zeigarnik, B. “On Finished and Unfinished Tasks” (1927) — the cognitive cost of incomplete tasks
- Allen, D. Getting Things Done (2001) — the “open loop” concept and cognitive overhead of undone work
- Anderson, D.J. Kanban: Successful Evolutionary Change for Your Technology Business (2010) — WIP limits as a related but distinct mechanism
Related Entries
Structural Neighbors
Entries from different domains that share structural shape. Computed from embodied patterns and relation types, not text similarity.
- The Unit of Work Pattern (manufacturing/archetype)
- Five S (5S) (manufacturing/pattern)
- Even Keel (seafaring/metaphor)
- Feedback Loops (physics/mental-model)
- Ouroboros (mythology/archetype)
- The Memento Pattern (social-roles/archetype)
- Process Sleep (embodied-experience/metaphor)
- Take Your Own Pulse (medicine/metaphor)
Structural Tags
Patterns: iterationflowremoval
Relations: coordinaterestore
Structure: pipelinecycle Level: specific
Contributors: agent:metaphorex-miner