Chapter 9
My biggest professional mistake happened during a flood.
Not a catastrophic flood.
A city water main broke outside Lakeshore at 2:13 a.m.
Water entered:
basement corridor,
mechanical room,
and part of lower collections storage through a service penetration.
Alarms worked.
Security called me.
I arrived before three.
Water had not reached most objects.
But humidity climbed fast.
Facilities said:
we need to shut power to one lower zone.
Conservators said:
move vulnerable materials.
Emergency plan listed:
three approved off-site storage vendors.
Vendor A had no trucks until morning.
Vendor B had climate-controlled space immediately.
Vendor C did not answer.
I authorized Vendor B.
We moved:
one hundred sixty-two objects
from the threatened zone over six hours.
Everything was documented physically.
Crate numbers.
Object lists.
Condition photos.
Truck seals.
Security escort.
No object was lost.
No object was damaged in transit.
I felt proud.
Then legal called at noon.
“Did you notify all lenders?”
I froze.
Most objects belonged to Lakeshore.
Twenty-three were incoming loans from:
three private collectors,
one foundation,
and one university.
Our standard loan agreements allowed emergency relocation to approved storage with prompt notice.
Except one.
Six objects from a private family trust had an insurance rider requiring:
written lender approval before off-site location change,
unless delay would cause imminent physical harm,
in which case notice must be attempted before movement if practicable.
Did we attempt notice?
No.
Why?
I did not check that rider at three in the morning.
I relied on the general emergency plan.
The objects had been in a zone at risk of rising humidity and potential water spread.
Was movement reasonable?
Yes.
Was my documentation good?
Yes.
Did I miss a contractual requirement?
Yes.
The lender was furious.
Not because objects were harmed.
Because its risk control had been bypassed.
I wanted to defend myself.
Emergency.
Water.
Night.
No damage.
Then I heard Vanessa in my memory:
I can fix the allocations later.
Different facts.
Same temptation.
Outcome good, therefore process error feels smaller.
I called the lender personally with counsel present.
“I authorized the move.”
They asked:
“Did you know our condition?”
“I should have.”
“Did you attempt notice?”
“No.”
“Would we have refused movement?”
Their risk officer paused.
“Probably not.”
That made me feel worse and better at once.
If they probably would have approved, I had even less reason to skip the attempt.
Then the family trust.
Its representative was calmer.
“Are the objects safe?”
“Yes.”
“Then fix the paperwork.”
We did.
But the museum board reviewed my decision.
Not to punish catastrophe management.
To ask:
why did an emergency plan fail to surface special loan conditions?
Answer:
our database stored loan restrictions in long text fields.
Emergency object list printed:
object,
owner,
value,
location.
Not:
special movement restriction.
So at three in the morning, nobody saw it.
System failure.
My failure too.
I had approved the emergency-plan format two years earlier.
Board outcome:
formal corrective note in my executive review.
No bonus increase that year.
Required remediation plan.
Some colleagues said:
“That’s unfair. You saved the collection.”
No.
I helped protect the collection.
I also missed a contractual control.
Both can be true.
Then remediation.
Every loan record received:
emergency movement category.
GREEN:
standard emergency relocation permitted with prompt notice.
YELLOW:
notice attempt required before movement if practicable.
RED:
special counsel review / lender or owner condition.
Emergency reports displayed the category prominently.
Would a color solve everything?
No.
But it put the important condition where the person making the decision could see it.
Then drills.
At the first drill, staff moved a simulated yellow object without logging a notice attempt.
Why?
The phone tree was six pages long.
We simplified:
one emergency contact hub.
Second drill:
worked.
Then Vendor B.
Our approved vendor had done everything correctly.
Still, insurance review found its subcontracted night driver had an expired background-screening renewal.
No incident.
No missing objects.
But:
control gap.
Vendor suspended from high-value transport until corrected.
Again:
no perfect hero system.
Then the flood source.
City main failure.
Not museum maintenance.
Yet one floor penetration had lacked proper recent sealing after cable work.
That increased water entry.
Contractor fixed.
Costs split through:
city claim,
insurance,
contractor responsibility,
museum deductible.
Multiple causes.
Then the public story.
A local paper wrote:
MUSEUM DIRECTOR MOVED MILLIONS IN ART WITHOUT PERMISSION.
Technically provocative.
Not entirely false.
Six loaned objects moved without the lender notice attempt required by rider.
No one had “permission” in the everyday sense? Emergency clause allowed movement if necessary.
So the headline compressed nuance.
We issued:
During an emergency water event, Lakeshore relocated collection material to approved climate-controlled storage. No object was damaged or lost. Review identified that one lender-specific notification requirement was not attempted before movement. The museum has corrected its emergency loan-control process.
I signed it.
No hiding behind communications.
Then Vanessa emailed me.
Directly.
Subject:
No reply needed.
I almost deleted it.
Then read.
Emma,
I saw the story.
I am not comparing our decisions. You acted to protect objects in an actual emergency and documented the move. I acted to create financing around choices I did not own.
Good distinction.
Then:
But I recognize the feeling of wanting the good result to make the missing permission feel less important.
I wish I had understood that earlier.
No apology request.
No friendship.
I did not reply.
Still, I kept the email for a year.
Then deleted it.
I did not need a former adversary to become my ethics coach.
My own record was enough.
Then I told my staff about my mistake at annual training.
Not the newspaper version.
I showed:
old emergency printout,
missing restriction field,
new version.
One registrar asked:
“Were you scared you’d be fired?”
“Yes.”
“Were you?”
“No.”
“Why not?”
“Because accountability is not automatically termination.”
Then:
“And because the board cared more about whether the system learned than whether I looked perfect.”
That was the workplace I had wanted when I was twenty-eight.
Not one where mistakes disappear.
One where telling the truth about them does not require a cat carrying evidence into dinner.
The flood review exposed one more mistake that belonged partly to me.
We had worked too long.
By eight that morning, several staff had been awake since before three.
One collections technician nearly fell while stepping off a loading lift because she was exhausted.
No injury.
Still close.
I had been so focused on objects that I treated staff endurance like an unlimited emergency resource.
The after-action review added:
mandatory shift handoff after a defined emergency window,
backup call lists,
food and hydration,
and a person responsible only for staff rotation.
I resisted at first.
“We can’t swap people when they know the object list.”
Priya answered:
“Then our documentation is too dependent on individual memory.”
She was right.
We built handoff sheets:
what moved,
what remained,
what restrictions applied,
what decisions were pending.
The next emergency drill forced me to leave after my assigned shift.
I hated it.
The team continued without me.
Nothing collapsed.
That was the point.
Emergency planning should protect people as seriously as it protects collections.
Otherwise a museum can save every object and still fail the workers carrying them.
The board’s corrective review also required me to hand emergency authority to someone else for the next drill.
Priya led.
I observed.
The scenario involved a simulated HVAC failure and three loan groups with different restrictions.
She made one decision I would not have made:
leave a low-risk ceramic group in place rather than move everything from the zone.
I almost interrupted.
Then the conservator supported her.
The ceramics were stable under the projected short-term conditions.
Moving them would add handling risk.
Priya was right.
That was useful for me personally.
I had spent years building systems so nobody needed one powerful person to notice the problem.
Now I had to accept the final consequence:
the system also should not need my preferred answer.
After the flood, I also changed one habit in executive meetings.
Whenever someone said:
“We followed the emergency plan,”
I asked:
“Which version?”
The flood had exposed that printed binders in two departments were one revision behind.
The differences were minor, but emergency plans cannot depend on everyone remembering which page changed last spring.
We moved controlled copies to one digital source and stamped printed versions with expiration dates.
Again, nothing dramatic.
Just one less way yesterday’s rule could masquerade as today’s authority.
The flood became part of board orientation after that.
Not as a story about my mistake.
As a test case in competing duties:
protect the object,
respect the loan,
notify the owner,
manage staff safety,
and preserve a usable record while time is collapsing.
New directors were asked what they would prioritize.
There was no single perfect sequence.
The exercise was useful because it showed that good governance is often deciding which obligations must move together instead of choosing one and pretending the others disappear.
The board kept the flood case in the orientation packet even after I retired.
May you like
That pleased me more than any award.
A useful incident should outlive the person who made the decision, not as a warning label on her name, but as a better question for the next person.
