DEV Community

Serguey Asael Shinder
Serguey Asael Shinder

Posted on

It Worked Is Not the Same as It Was a Good Call

Somebody deployed straight to production
on a Friday evening,
no review,
because a customer was shouting.

It worked.

On Monday they were thanked
in front of everyone.

The same month
somebody else took a week longer
to ship a change properly.
Behind a flag,
with a rollback written,
tested against real volumes.

A dependency they could not have seen
failed halfway through the rollout,
and the failure
is what people remember.

Look at what the team just learned.

Not from the handbook.
From the applause.

The shortcut pays.
Care is a risk to your reputation.

Nobody said it.
Nobody had to.

This is the most ordinary mistake
in how decisions get judged.

We grade the choice
by how it turned out,
because the result is visible
and the reasoning is not.

But a result is the decision
plus luck,
and you only ever see
the one version of events
that happened.

The Friday deploy
might have taken the site down
two times in three.

You will never know,
because the one run you can see
was the good one.

It does the same thing
to your own habits.

Every risky call that lands
gets filed in your memory
as judgement.

Every careful call that goes wrong
feels like proof
that care is slow and pointless.

Give it a few years
and you are a person
who takes big risks
with great confidence
and has no idea
how many times you were rescued.

So judge a decision
by what was known
when it was made.

What did they know.
What could they reasonably have found out.
What would finding out have cost.
What was the downside,
and who would have carried it.

Then ask one thing
about the result.

With the same information,
knowing it could go either way,
would we make that call again?

If not,
it was a bad call that got lucky,
and it deserves a quiet word,
however well it went.

If so,
it was a good call
that ran into the world,
and it deserves defending,
out loud,
by somebody senior enough
to make the defence stick.

Luck visits everyone.

Do not let it
write your standards.

– Serguey Asael Shinder

Top comments (2)

Collapse
 
unitbuilds profile image
UnitBuilds •

Do not follow external links, this is a phishing scam. DEV.to uses Sloan for automated messaging.