Sprint planning.
You take the ticket,
open it up in your head,
and walk through the steps.
Add the field.
Change the API.
Update the screen.
Tests.
Three days.
It feels honest,
because you can see every step.
That is the problem.
Your last six tickets of this size
took between five and nine days.
You know that.
If someone asked,
you could tell them why.
The review that sat for two days.
The migration nobody mentioned.
The flaky test.
Each one feels like bad luck,
a thing that happened once.
So none of them
make it into the next estimate.
In 1994
Buehler, Griffin and Ross
asked final year students
when they would finish their thesis.
On average
they said thirty four days.
On average
it took fifty five.
Fewer than a third
finished by their own date.
The researchers also asked
for a date if everything went
as badly as it could.
The average real finish
came later
than the average worst case.
This is the planning fallacy,
and it is not ignorance.
The students knew
their past work had run late.
They explained it away,
the way you do,
as special.
Daniel Kahneman
called the cure
the outside view.
Stop asking
how this task will go.
Ask how tasks like it
usually went.
He told the story on himself.
His team writing a school curriculum
guessed about two years.
An expert in the room
knew that similar teams
had taken seven to ten,
and four in ten never finished.
It took eight.
So keep the numbers.
For every ticket,
the estimate
and the actual,
one column each.
It takes a minute.
After a month
you have a reference class,
your own,
and it is far more honest
than your imagination.
When you estimate,
start from that.
Five to nine days for this size.
Then adjust,
a little,
for what is truly different.
Not for what feels different.
Every task feels different
from the inside.
That is what inside means.
The plan in your head
has no reviews waiting,
no surprises,
no lost Tuesday afternoons.
The record does.
Believe the record.
– Serguey Asael Shinder
Top comments (0)