Pull-Based Project Method
Limit active projects and pull new work only when capacity becomes available.
- Difficulty
- Moderate
- Time to result
- ~weeks to results
- Steps
- 6
- Confidence
- 99%
The Pull-Based Project Method limits work at two levels. During a work period, the brain focuses on one thing rather than switching among tasks. Across the broader project portfolio, only two or three projects remain active. Everything else sits in a holding tank and creates no personal obligation until capacity opens. When one active project finishes, the worker pulls the next item from the queue. This prevents the overhead tax generated by every live commitment: emails, meetings, planning, status checks, and the cognitive burden of remembering unfinished work. Without a limit, these taxes compound into an overhead spiral where maintaining projects consumes nearly all available capacity and completion slows further. The method changes assignment from a push system—distribute everything now—to a capacity-aware pull system.
Origin
Extracted from Deep Dive with Ali Abdaal during a discussion of applying software-development pull systems to broader knowledge work.
Core principles
- 01Work on only one object of attention at a time.
- 02Keep roughly two or three projects active overall.
- 03Store future work in a holding tank rather than assigning it early.
- 04Pull new work only after an active project finishes.
- 05Every active commitment creates administrative and cognitive overhead.
How to run it
- 1
Inventory Live Commitments
List every project currently considered assigned, including work that is waiting on others but still generates meetings, email, or worry.
Pro tip Count obligations, not just projects receiving hands-on effort today.
- 2
Set the Active Limit
Choose approximately two or three projects to remain active. Base the exact limit on real administrative and cognitive capacity.
Pro tip Start with three and reduce further if progress remains fragmented.
Watch out Scheduling each of fifteen projects on a different day does not remove their overhead.
- 3
Create a Holding Tank
Move all other possibilities into a trusted backlog where they are visible but not assigned to an individual.
Pro tip Record enough context to reconsider the item later without maintaining it now.
Watch out Do not continue holding standing meetings for backlog items.
- 4
Single-Task the Work Block
Within each work period, make decisions and progress on one active project before moving to another.
Pro tip Close email and unrelated project channels during the block.
Watch out Frequent switching forces the brain to rebuild its internal model.
- 5
Pull Only After Completion
When a project leaves the active set, reassess the holding tank and deliberately pull one replacement.
Pro tip Use a regular planning review to choose the next item by current value.
Watch out Do not refill capacity merely because a project pauses briefly.
- 6
Monitor the Overhead Tax
Track how much time active projects generate in meetings, communication, planning, and mental maintenance. Lower the active limit if overhead starts crowding out substantive work.
In the wild
A company has ideas for website summaries, Patreon, a keyboard, a bag, and many other initiatives. Instead of distributing all of them to team members, it keeps them in a bucket list, activates only the current six-week priorities, and revisits the rest later.
→ Team members stop treating deferred ideas as current obligations and recover cognitive capacity.
Common mistakes
Pushing Every Task onto Someone's Plate
Assignment makes a project psychologically and administratively live even when no substantive work is underway. Keep future work unassigned in the holding tank.
Counting Only Hands-On Work
A waiting project still produces messages, meetings, planning, and worry. Include those costs when setting the active limit.
Pulling Before Capacity Is Truly Free
A temporary pause is not completion. Pulling replacement work too early recreates excessive work in progress.
Is it for you?
Best for
Knowledge-work teams and individuals carrying many possible projects or loosely assigned responsibilities.
Not ideal for
Roles requiring genuine simultaneous incident response across several critical systems.
From the transcript
“I'm a big believer in like pull-based methodology where there's like two or three things you're two or three things you're working on.”
“When something finishes, you can pull something else in.”
“I honestly think like three active projects at a time is best.”
From the episode
7 Unconventional Insights To Level Up Your Productivity