Task Granularity: Finding the Line Between Too Fine and Too Coarse

One team’s board has a task called “launch the product.” It’s been sitting in the in-progress column for six weeks. Nobody can tell from looking at it whether it’s a day from done or a month away, because a task that broad carries no information about where the actual work stands. Another team, two floors up, has a board with four hundred cards, and half of them read like “update button color” and “fix typo in footer.” Their board is technically precise and completely unreadable, because finding the three cards that actually matter this week requires wading through a swamp of administrative noise.

Both teams have a task management problem, and it’s the same problem wearing opposite costumes. Granularity — how big or small a unit of work should be before it earns its own task — is one of the least discussed decisions in how teams actually organize work, and getting it wrong in either direction quietly breaks the same thing: the board’s ability to tell anyone, at a glance, what’s actually happening.

What Too Coarse Actually Costs

A task that’s too broad hides risk. “Launch the product” sitting in progress for six weeks could mean everything is on track and the work is simply large, or it could mean three separate blockers have piled up that nobody’s surfaced because there’s no smaller task where a blocker could be attached and made visible. Coarse tasks compress a lot of actual state into a single status field, and a single status field — todo, in progress, done — has nowhere near enough resolution to represent six weeks of shifting reality. The task isn’t lying, exactly. It just isn’t saying enough to be useful to anyone who isn’t already deep in the work.

What Too Fine Actually Costs

The opposite failure is less obviously a failure because it looks like diligence. Breaking work into extremely small tasks feels thorough and gives a satisfying sense of progress every time a card moves to done. But a board with hundreds of tiny tasks buries the handful that actually carry risk or decision points under a pile of routine, low-stakes items that didn’t need tracking as individually as they got tracked. Finding out what’s actually at risk this week means scanning past dozens of cards that were never going to be interesting, and the cognitive cost of that scanning adds up, especially for anyone — a manager, a stakeholder — who isn’t in the work daily and relies on the board itself to orient them quickly.

A Rough Test for the Right Size

There’s no universal correct task size, because it depends on how much uncertainty and coordination a piece of work actually carries. But a useful working test: a task is sized about right if, when it’s sitting in “in progress” for longer than expected, that fact alone would prompt a specific, useful question from someone looking at the board. “Launch the product has been in progress for six weeks” doesn’t prompt a useful question — the task is too broad to know what to even ask. “Update button color has been in progress for six weeks” would prompt a useful question, but the wrong one — why is a five-minute task taking six weeks, probably because it’s stuck behind something bigger that isn’t itself being tracked.

Task size signal	Likely problem
Sits in progress for weeks with no sub-visibility	Too coarse — hiding real state
Takes under 30 minutes but has its own card	Possibly too fine — adds noise without adding signal
Stalling prompts a specific, answerable question	Reasonably well sized
Board has more cards than anyone can scan in two minutes	Too fine on average, or too much routine work tracked individually
Nobody can guess remaining effort from the task title	Too coarse

Different Kinds of Work Warrant Different Granularity

The mistake many teams make is applying one granularity rule uniformly across all their work, when actually the right level of detail depends on what kind of work it is. Routine, well-understood work — the kind a team has done dozens of times before — can usually be tracked coarsely, because there’s little uncertainty to surface; everyone already knows roughly what’s involved. Novel or high-risk work benefits from finer granularity precisely because uncertainty is high, and finer tasks give more places for a blocker or a changed assumption to attach itself and become visible early rather than staying buried inside one large, opaque task until the deadline arrives.

Breaking Tasks Down as You Learn More, Not All at Once

One practical habit: rather than trying to decide the perfect granularity upfront, start a large task coarse and split it the moment real uncertainty or a dependency surfaces inside it. “Launch the product” might reasonably start as one task during early planning, when the actual steps aren’t yet clear. The moment it becomes apparent that legal review, a partner integration, and a marketing asset all need to happen somewhat independently, that’s the signal to split it into three tasks — not because a rule says tasks should always be split that finely, but because the underlying work has revealed enough distinct, independently-trackable pieces to warrant it.

Letting the Board Serve Its Actual Audience

Part of what determines the right granularity is who’s actually reading the board and what they need from it. A board meant primarily for the people doing the work day to day can tolerate more detail, because they have the context to filter it quickly. A board that stakeholders or leadership check periodically needs a coarser layer — often a rollup view rather than the working team’s actual task list — because forcing an occasional viewer to parse four hundred granular cards to find the three that matter to them is asking them to do work the board should be doing instead.

Revisiting Granularity as the Work Changes

Granularity isn’t a decision made once at project kickoff and left alone. As a project moves from planning into execution, and then into a stabilization or wrap-up phase, the right level of detail shifts too — early uncertainty calls for finer breakdown, later routine execution can consolidate back into coarser tracking without losing anything important. Teams that periodically ask whether their current task size is still telling them anything useful catch the drift toward either extreme well before it turns into a board nobody can actually read at a glance.