Software used to show you
what it could do.
A menu.
A row of buttons.
A settings page you could read
top to bottom
while avoiding the actual work.
That was not clutter.
That was an inventory,
and it was a promise.
Everything this thing does
is visible somewhere,
and you can find it
by looking.
Now the interface is a box
that says ask me anything.
An empty box has no edges.
To ask for something,
you have to already believe
it is possible,
and nothing on the screen
is telling you what is.
So people guess,
narrowly,
based on what they expected
before they arrived.
They ask for the two things
they came in with,
get them,
and the other forty things you built
are never found,
because there was no surface
on which to find them.
Watch what that does
to everyone involved.
Your users settle at a fraction
of the product
and call that the product.
Your support team spends its life
answering yes, it can do that.
And you lose your instrument.
You can no longer tell
whether a feature failed
or was simply never discovered,
because both look
like nobody using it.
The knowledge of what the thing does
moves out of the software
and into folklore.
Word of mouth.
A colleague who shows you.
A thread somewhere.
That is a fragile place
to keep your product.
So as more of this gets built,
the scarce skill stops being capability
and becomes legibility.
Showing the shape of what is available.
A short menu of what people actually do.
Examples that are real,
not charming.
A reply that says
what else you could have asked,
in this situation, now.
Discovery used to arrive free,
paid for by the layout.
Now somebody has to design it
deliberately,
and mostly nobody has,
because the box looked like
it had solved that.
Put a visible surface
on the invisible thing.
What people cannot see,
they cannot ask for.
What they cannot ask for,
you did not build.
– Serguey Asael Shinder
