It was the worst month
the team had had.
Four incidents.
A rollback on a Friday.
A customer escalation
that reached a director.
So you acted.
A new rule
that every change to the billing code
needs two approvals
and a written test plan.
The next month
there was one incident.
The rule stayed.
It is still there,
three years later,
and everybody knows why.
It worked.
Maybe.
Any number you measure about a team
is part skill
and part luck.
Incidents.
Bugs found after release.
Sprint results.
Survey scores.
When a number lands far from its usual place,
part of the reason
is usually luck,
and luck does not repeat on schedule.
So the month after an extreme one
tends to be less extreme,
whether you do anything or not.
This is regression to the mean,
and it makes action
look effective
and praise look dangerous.
The classic story is about
flight instructors.
They found that pilots who were praised
after a great landing
usually did worse the next time,
and pilots who were shouted at
after a bad one
usually did better.
So they concluded
shouting works.
It did not.
The landings were drifting
back to normal
on their own.
You are standing exactly where they stood.
You react to extremes,
because extremes are what get noticed.
So every intervention you make
gets measured at the bottom of a dip,
and the climb back
is put down to you.
Do it for a few years
and you collect
a lot of rules,
and a quiet belief
that being hard on people
gets results.
So before you react,
look at the longer line.
Twelve months, not one.
Was this month
outside anything that has happened before,
or just the low end
of a normal spread?
When you do change something,
write down first
what result would show it worked,
and how long you will wait.
Compare with a team
that did not change.
Remove a rule now and then
and see if anything happens.
And be slowest
to draw lessons about people.
A good month after a hard talk
proves very little.
Neither does a bad month
after a compliment.
Most numbers
wander back home.
Do not take the credit
for the walk.
– Serguey Asael Shinder
Top comments (0)