Executive summary
Being busy all the time isn’t the same as making progress
The essentials of the report, to read in two minutes. The full report, with every number, its sources and the studies, is a PDF download.
The report has 18 pages and weighs 3 MB. You can share it as is, crediting the author.
Mosca del WIP is a one-minute game. You choose how many tasks an office fly can have started at once, from 1 to 10, and you watch her workday in front of her board. That number is her limit on work in progress, or WIP.
Inside there is a computer experiment, with the real map of a fruit fly’s brain and office rules inspired by studies with real people. The computer played 200 days with each limit and we took the average.
What did the fly gain by limiting her WIP?
She finishes more, and each task is done sooner.
with 2 tasks at once
18 finished per day, 1 h each
with 9 tasks at once
9 finished per day, almost 9 h each
Work moves forward instead of waiting.
with 2 tasks at once
each task moves forward almost half the time
with 9 tasks at once
each task waits 9 out of every 10 minutes
Less overload.
with 2 tasks at once
she ends the day Calm, without smoking
with 9 tasks at once
she ends the day Overwhelmed, with 13 cigarettes
In the game, average of 200 days played.
What we found
One task at a time wasn’t the best
In the game, with 1 task at once the fly finishes 15.9 per day, fewer than with 2. If that task is left waiting on someone, she has nothing to do. With up to 3 she still ends the day Calm, and that is her sustainable limit.
What’s started doesn’t disappear
In the game, the fly starts with 10 tasks at once and at 12:00 her limit drops to 2. She takes no new tasks, and only at 14:35 does she have 2 open again. Lowering the limit doesn’t calm her right away.
Each urgent task cost one regular task
In the simulation, the fly gets urgent tasks that jump the limit. With a limit of 2, each urgent task per day costs her about one regular task. A limit of 2 is still the one where she finishes the most.
Studies with real people point in the same direction and none of them gives a number. Your team’s number comes from its own test, and the four steps to run it are further down.
What studies with real people say
In Italy, judges who were assigned more cases at once by lottery took longer to close them, and the quality of their rulings didn’t change (Coviello, Ichino and Persico, 2014). Recruiters finished more with a bit of parallel work, and past a point they finished less (Aral, Brynjolfsson and Van Alstyne, 2012). DORA’s surveys of software teams say that WIP limits help, especially with a visible board. None of them measured how many things it’s best to have open.
What comes from the game and what doesn’t
The game is a single fly on a single day, with no coworkers to hand work to, and it doesn’t measure people or teams. Little’s law, which is math, says that, at the same output rate, each task takes longer when more are open. But that she finishes half as many with 9 depends on two rules of ours, the cost of switching tasks and the cigarettes, and that stress goes up is our assumption. Her brain is real, but it doesn’t decide anything.
What your team takes away
It’s what stays in plain sight after trying a limit for two weeks. Your team’s own test will show how much it gains.
-
Their load in plain sight, how many things each person has started.
-
Where the work sits waiting.
-
An agreed limit, written on the board.
-
Their own data after two weeks.
Four steps to try a limit with your team
The game doesn’t tell you your team’s number. These four steps help you find it.
Make the work visible
Three columns are enough (To do, In progress, Done). What you can’t see, you can’t limit.
Count what’s open and set a limit
Count how many things each person has open today. Agree on a low limit, write it on the board and stick to it. If some waiting can’t be avoided, make it more than one.
When someone is left waiting, go after the wait
Don’t open something else to “make good use of the time.” Help unblock what’s waiting. With fewer waits, the limit can go down.
Try it for two weeks and measure
On each card, note the day it started and the day it finished. After two weeks, compare how many got finished, and how long each took, with the two weeks before.
A sustainable limit, not one thing at a time.
What’s in the report
The report is a PDF of 18 pages and takes about 10 minutes to read. Here is what’s in it, page by page.
- p. 1Why you’re seeing this Busy all day. So what did you finish?
- p. 2The star and her rules A real fly in a make-believe office, with rules you can see
- p. 3The same day, twice With 9 tasks she was busy all day and finished half as many
- p. 4From 1 to 10 tasks There’s no number where everything breaks. Each extra task costs a little more
- p. 5The paradox With more tasks she never runs out of things to do, but each one spends more time waiting
- p. 6The limit Your ideal limit depends on how much you wait on others
- p. 7What’s started doesn’t disappear Lowering the limit doesn’t calm her right away. What’s started takes over 2 hours to clear
- p. 8With real people It’s not just the fly. Judges, recruiters and teams point in the same direction
- p. 9The disguise The fly doesn’t smoke. The cigarette is a disguise for cleaning her antennae
- p. 10The brain This is how her brain lights up with 2 and with 9 tasks
- p. 11Urgent tasks With urgent tasks the best limit is still 2, and each one costs a regular task
- p. 12What to do with this Four steps to try a limit with your team
- p. 13Appendix A, is it real? What’s math, what comes from studies and what we added
- p. 14Appendix B, the numbers Changing the rules barely moves the best limit, but it does change how much output drops
- p. 15Appendix C, sources and credits Where each idea comes from
The report has 18 pages and weighs 3 MB. You can share it as is, crediting the author.