Habits & Routines · Career

Why Boards Beat Lists: Kanban, WIP Limits & Little's Law

A to-do list tells you what exists; a board tells you where work is stuck. The math behind that difference is Little's Law — and it says the fastest way to finish more is to start less.

By Hank · · Editorial standards

https://taskcoach.ai/blog/kanban-wip-limits-littles-law/

The List Lies By Omission

A to-do list answers one question: what exists? It cannot answer the questions that actually determine whether you finish anything: What's in motion? What's been in motion for three weeks? Where is work piling up?

A kanban board answers all of them at a glance, because the board's columns encode state — Backlog, Todo, In Progress, Done — and cards physically accumulate where work is stuck. Four cards jammed in one column is a bottleneck diagnosis you didn't need a meeting for.

That visibility is nice. But visibility isn't why kanban makes you faster. Math is.

Little's Law: The Uncomfortable Equation

In 1961, MIT's John Little proved a deceptively simple queueing-theory result that holds for any stable system — assembly lines, supermarket queues, hospital wards, your project board:

Average cycle time = Work in progress ÷ Throughput

Your throughput — how many things you actually complete per week — is roughly fixed. It's set by your hours, energy, and skill, and it doesn't change because you're excited. Which leaves exactly one variable under your direct control: how many things you allow in progress at once.

Run the numbers. You finish about one meaningful item per week:

  • 6 items in progress → each takes an average of 6 weeks to complete
  • 2 items in progress → each takes an average of 2 weeks

Same person. Same effort. Same throughput. Tripling of speed per item, purchased entirely by starting less.

Every card you start joins a queue that everything else has to wait behind.

The Psychology Agrees With The Math

The queueing argument alone understates the case, because high WIP doesn't just divide your capacity — it shrinks it.

The switching tax. Rubinstein, Meyer & Evans (2001) measured the cost of shifting between tasks: performance drops and the "resumption lag" of reloading context can consume 20-40% of productive time for complex work. Six parallel projects means paying that toll dozens of times per week. The capacity you're dividing is smaller because you divided it.

The open-loop tax. The Zeigarnik effect — Bluma Zeigarnik's 1927 finding that interrupted tasks stay mentally active — means every In Progress card keeps a background process running in your head, including at dinner. Masicampo & Baumeister (2011) showed these open loops measurably impair performance on unrelated tasks — and that the interference stops when the task has a concrete plan. A board with a small In Progress column is, cognitively, a shorter list of things your brain refuses to stop rehearsing.

Kanban's Origin: The Supermarket Insight

Taiichi Ohno built the original kanban system at Toyota in the 1950s after studying American supermarkets: shelves restock only when a gap appears — a pull system, downstream demand pulling upstream work. The result was the Toyota Production System, the most-studied manufacturing method in history.

Personal kanban (formalized by Jim Benson and Tonianne DeMaria Barry in Personal Kanban, 2011) ports exactly two rules from all that machinery:

  1. Visualize the work.
  2. Limit work-in-progress.

That's the whole system. Everything else is decoration.

Pull, don't push: a card advances only when capacity opens downstream.

Running It Without Fooling Yourself

1. Make Backlog vs Todo a real border. Backlog is deliberately parked and guilt-free; Todo is committed. The move from Backlog to Todo should happen only when something finishes — that's the pull system. If everything enters at Todo, you've rebuilt the infinite list with extra columns.

2. Cap In Progress at 2-3. When you want to start something new and the column is full, the rule triggers the only two honest options: finish something, or consciously demote something back to Backlog. Both are wins. Silent parallel expansion is the loss.

3. Read the pile-ups, not the cards. The board's diagnostic power is in the distribution. Cards accumulating in one column = bottleneck. A card that hasn't moved in two weeks = it's mislabeled (too big, blocked, or secretly a Won't-do). Act on the pattern.

4. Let moving a card back be a power move. Demoting work to Backlog is the WIP limit functioning as designed — a decision that frees capacity — not an admission of defeat.

What TaskCoach.AI Does With This

Every Space ships with a Kanban view over the same task database as the other seven views — dragging a card to Done stamps the completion, pays the XP, and feeds your analytics, so the drag is the data entry. Backlog and Todo are separate columns on purpose: the parked/committed distinction is the WIP limit's load-bearing wall. And custom columns let you match your real pipeline (Scoping → Drafting → Review → Shipped) so the bottleneck shows up where it actually lives.

The Bottom Line

Your completion rate is fixed. Your WIP is a choice.

Little's Law turns "focus" from a virtue into arithmetic: every item you start joins the divisor under everything else. The board isn't a prettier list — it's the instrument that makes the divisor visible.

Stop starting. Start finishing.

Frequently asked questions

What is Little's Law?

A queueing-theory result proved by John Little in 1961: the average number of items in a system equals the arrival rate multiplied by the average time each item spends in the system. Rearranged for personal work: your average cycle time equals your work-in-progress divided by your completion rate. It holds for any stable system — factory lines, coffee queues, and your project board.

Why do WIP limits make you faster? Isn't that backwards?

It feels backwards because starting things feels like progress. But your completion rate is roughly fixed by your hours and energy — so the only lever left is how many things share that fixed capacity. Six items in progress at a rate of one finished per week means every item averages six weeks. Two items in progress means two weeks. Same effort, three times faster per item, plus compounding gains from less context-switching.

What's the difference between Backlog and Todo?

Commitment. Todo means committed — it consumes attention and counts against your working set. Backlog means deliberately parked with zero guilt attached. A healthy board moves items from Backlog to Todo only when capacity opens up. Moving an item back to Backlog is a decision, not a failure — it's the WIP limit doing its job.

How many things should be In Progress at once?

For an individual: one to three. The research on task-switching costs suggests every additional parallel stream taxes the others; most personal-kanban practitioners (Benson & Barry's Personal Kanban recommends starting at three) converge on 2-3 as the ceiling where flow survives. If your In Progress column is regularly above five, it's a second Backlog wearing a costume.