“Build a website,” “move,” “write a thesis,” and “find a job” look like tasks, but they are projects. Each contains decisions, dependencies, and unknowns.
When a project is stored as one line, you must rediscover what to do every time you return. When it is split into a hundred tiny items, the list becomes another source of overload.
Useful decomposition sits between those extremes.
Start with the outcome, not the task list
Write down what should exist or change when the project is finished.
Vague:
Work on my portfolio.
Verifiable:
Publish a page containing three projects, an explanation of my role, and contact details.
The outcome limits the project. Without it, you cannot tell which actions are necessary and which merely look productive.
Identify intermediate outcomes
Do not jump directly to micro-steps. First find three to seven larger states.
For the portfolio:
- three projects are selected;
- the source materials are collected;
- the case descriptions are written;
- the page is assembled;
- everything is checked and published.
Each stage should leave a new object or decision. “Work on the website” describes activity, not an intermediate result.
Find the current stage
Not every part of the project deserves attention at once. Ask:
Which intermediate outcome unlocks the next one?
If you have not selected the projects, choosing page typography is premature. If the report requirements are unknown, polishing paragraphs is risky.
Keep the current stage and a few nearby actions visible. The complete map is for orientation, not daily pressure.
Turn the stage into observable actions
Use this formula:
verb + object + boundary
For example:
- open the folder of previous projects;
- list five candidates;
- rate them against three criteria;
- select three projects;
- record the missing materials.
Words such as “develop,” “advance,” and “figure out” are warning signs when no physical action is visible behind them.
Test the size of a step
A step is ready when it:
- makes sense without more planning;
- contains no hidden decision;
- fits in one work session;
- has a visible completion condition;
- uses materials you can access.
“Write the section” may work for a familiar topic. For an unfamiliar one, a better step might be “collect five sources and list the open questions.”
Record dependencies separately
Large projects often stall because of waiting:
- for a colleague’s response;
- for a document;
- for payment;
- for another person’s decision;
- for access to a system.
Track the dependency as its own item:
| Waiting for | Owner | Follow-up date | Parallel work |
|---|---|---|---|
| Sales data | Analyst | Wednesday | Build the report outline |
“Waiting for information” does not need to freeze the entire project.
Do not plan uncertainty with false precision
Later actions in a new project will change. Create:
- a detailed plan for the nearest stage;
- a rough map of later stages;
- a list of open questions;
- a review date.
This rolling plan is more honest than eighty detailed items, half of which will become irrelevant.
When decomposition becomes procrastination
Warning signs include:
- repeatedly changing the structure without producing an outcome;
- selecting another planning app instead of doing the work;
- detailing steps that cannot yet be predicted;
- spending longer formatting the plan than completing the first stage;
- still not knowing what to do now.
Set a timer for planning and finish with a substantive action. After twenty minutes of planning, send one request, create the first file, or test one assumption.
Example: write an expert article
Finished outcome:
an article answering one reader search intent is published
Stages:
1. intent defined
2. sources collected
3. outline completed
4. draft written
5. review completed
6. article published
Current stage:
define the intent
Nearest actions:
- list five versions of the query
- inspect the search results
- choose one central reader question
- write the article boundariesMove the project into Flammy
Do not create a daily task called “finish the project.” Choose a repeatable action that advances the current stage:
- one work block on weekdays;
- complete one nearest action;
- leave the next step before ending the session.
Update the wording when the stage changes. XP represents repeated work on the process, not a fictional percentage of total completion.
If the main difficulty is the instant of beginning, use the guide to starting a hard task.
Decomposition template
Finished outcome:
Intermediate outcomes:
1.
2.
3.
Current stage:
Nearest actions:
1.
2.
3.
Dependencies:
Next review date:Decomposition is not a promise of complete control. It should make the nearest section clear enough to enter.
For more causes and situations, see Procrastination: what to do.