Chapter 6
The spreadsheet appeared nine months after Darren left.
Not because anyone was hiding it.
Because nobody had asked for old pilot files.
Bellaforte had tested a revenue-planning tool two years earlier.
The vendor called it SeatFlow.
The software estimated likely check size based on:
party size.
reservation time.
past spend.
wine history.
special-occasion notes.
booking source.
It did not use age.
It did not know whether a guest looked wealthy.
It predicted from restaurant data.
The pilot lasted six weeks.
Vivian remembered it.
“Darren hated it.”
Priya raised an eyebrow.
“Why?”
“He said it underestimated walk-ins and regulars.”
Then IT restored an archived export.
Darren had created his own companion spreadsheet.
Not part of vendor.
Columns:
Table.
Reservation.
Expected spend.
Turn risk.
Bar likelihood.
Priority.
Then a final column:
NOTES.
Some notes were ordinary.
Birthday.
Client dinner.
Wine collector.
First date.
Regular.
Needs quiet.
Others were not.
Senior pair.
Solo reader.
No alcohol.
Tea only.
Old regular — sits.
Tourist family.
Young four-top.
Industry.
Emma stared when Priya showed her.
“I’ve seen this.”
“Where?”
“Printed at the host stand sometimes.”
Not every night.
Darren called it floor map.
Hosts thought it was planning.
Then the codes.
Green:
protect.
Yellow:
normal.
Red:
convert if waitlist.
Convert meant:
move to bar.
offer lounge.
speed check.
release incomplete party.
Who was often red?
Low predicted spend.
That included:
solo diners.
non-drinkers.
older regulars with known habits.
families sharing dishes.
guests with one-course history.
Not age alone.
But “senior pair” appeared in notes twenty-eight times.
Priya looked at Vivian.
“Who authorized these notes?”
“No one.”
Then:
“Did hosts enter age?”
Emma shook head.
“Darren did, mostly.”
Sometimes assistant managers.
Why?
To predict stay length.
That made age part of operational decision.
Not necessarily illegal automatically.
But serious.
Then Rose’s reservation.
Archive copy existed.
MANCINI.
Party 2.
Expected spend blank because standing reservation.
Turn risk:
HIGH.
Note:
senior woman arrived solo.
Priority:
RED — convert 8:15 if second absent.
Emma felt sick.
Darren had entered it at 7:48.
Before confronting Rose.
Before pretending his concern was only incomplete party.
He had decided she was a low-value table.
Then something else.
At 8:12, Darren changed expected spend manually:
$65.
Why $65?
Tea.
Maybe one entrée.
At 8:17:
waitlist party of four projected $780.
Then 8:19:
Rose “departed.”
The lie now had economic trail.
Leo read.
His face went still.
Mara said:
“Do not overstate.”
“What is there to overstate?”
“This does not prove he targeted your mother because she was old.”
Leo looked at senior woman arrived solo.
Mara continued:
“It proves he used age-related description as one factor in a revenue decision.”
Important.
Then other records.
A thirty-two-year-old solo diner with no alcohol history:
RED.
A sixty-eight-year-old couple with high wine spend:
GREEN.
So money dominated.
Age influenced turn-risk assumptions.
The system was not simple age discrimination.
It was economic profiling with age-coded shortcuts.
That nuance mattered.
Then Vivian found an old presentation.
Darren had shown the spreadsheet at manager meeting.
Slide:
Protect high-value seats.
Convert low-productivity tables.
Use lounge aggressively.
Vivian had attended.
Her notes:
Smart. Need consistency.
She closed eyes.
“I saw this.”
Priya said nothing.
“I didn’t see senior notes.”
“Did you see convert low-productivity?”
“Yes.”
“Did you ask what low-productivity meant?”
“No.”
There.
Again.
Then Leo.
Had he seen?
No.
But he had pushed throughput.
Vivian had praised Darren.
The culture was not secret enough to let ownership pretend innocence.
Then investor meeting.
One partner said:
“You are turning basic revenue management into moral scandal.”
Leo answered:
“Revenue management is fine.”
Then:
“Falsifying guest status is not.”
Partner:
“Agreed.”
“Using age as shorthand for who can be pushed?”
Silence.
Then:
“Also not fine.”
Another investor said:
“We do this everywhere. Families sit. Older couples linger. Business dinners spend.”
Priya interrupted:
“Operational patterns can be studied. The question is whether individuals are treated as predicted category before their behavior supports it.”
That quieted room.
Then SeatFlow vendor.
Bellaforte contacted.
Vendor confirmed:
no age field.
no demographic scoring.
no instruction to create manual age tags.
Darren’s spreadsheet was local.
Good.
No blaming algorithm.
Then current system.
Should Bellaforte stop using past spend?
They still knew regular guest history.
Wine preferences.
allergies.
favorite tables.
past check.
Personalization is hospitality.
But using spend to decide whether confirmed guest deserves table?
Different.
New rule:
Historical spend may inform service preparation and inventory.
It may not be used to cancel, reseat, or shorten a confirmed reservation without ordinary operational reason.
Then waitlist priority.
First available by reservation/waitlist rules.
No projected-spend sorting.
VIP contracted events separate.
Transparent.
Then Emma asked:
“What about somebody who spends five thousand every Friday?”
Vivian smiled.
“We know them.”
“Do they get better table?”
“Sometimes.”
Emma looked.
Vivian continued:
“If we choose to reserve a table for a regular before others book, that is a business choice.”
Then:
“But we do not take your table after you arrive because someone richer walked in.”
That line became useful.
Not equality of every perk.
Reliability of confirmed promise.
Then Darren’s prior settlement.
Could Bellaforte reopen because spreadsheet showed more intent?
Legal matter closed.
No.
Maybe new evidence could matter depending settlement language, but counsel saw no practical need.
Leo hated.
“He gets away with this?”
Mara looked.
“He was fired. Repaid money. His record reflects termination. What additional outcome are you seeking?”
Leo did not know.
Then Rose heard about spreadsheet.
At dinner.
She listened.
Then:
“He wrote senior woman?”
“Yes.”
“Accurate.”
Leo stared.
“Ma.”
“I’m senior.”
“That’s not point.”
“No.”
She smiled.
Then:
“The point is he thought he knew what I was worth before I ordered.”
Exactly.
Then she asked:
“Did he write difficult mother?”
Leo laughed.
“No.”
“He should have.”
Good.
Then Rose did something Leo did not expect.
She asked to see spreadsheet.
Mara initially hesitated because other guest info.
They redacted.
Rose reviewed her row.
MANCINI.
senior woman arrived solo.
RED — convert.
She stared for a while.
Then:
“This is uglier than him telling me to leave.”
Leo looked.
“Why?”
“Because telling me to leave was rude.”
She tapped row.
“This is calm.”
There.
No anger visible.
Just a system someone built to make rude treatment look like management.
Then Rose handed back.
“Do not keep showing me.”
“You wanted to see.”
“I saw.”
Enough.
Then Bellaforte deleted no records.
Historical pilot retained under policy.
Not public.
Not framed.
Then a current manager asked:
“Should we remove all guest notes about age?”
Mostly yes unless operationally relevant and lawful, such as a guest explicitly requesting accessibility? Better not note age; note needs.
New standard:
Record needs, not assumptions.
Instead of:
elderly guest.
Write:
requests chair with arms.
needs low-noise table.
prefers slower pacing.
or nothing.
That change seemed small.
It was not.
Words shape what the next employee thinks they already know.
Priya also reviewed whether any guest note using age-related language had been copied into permanent guest profiles.
Most had not.
Where it had, Bellaforte replaced assumption-based labels with specific service preferences or removed them entirely.
A guest file did not need to remember that someone “looked old.”
Priya also reviewed whether any guest note using age-related language had been copied into permanent guest profiles.
Most had not.
Where it had, Bellaforte replaced assumption-based labels with specific service preferences or removed them entirely.
A guest file did not need to remember that someone “looked old.”
Priya also reviewed whether any guest note using age-related language had been copied into permanent guest profiles.
Most had not.
Where it had, Bellaforte replaced assumption-based labels with specific service preferences or removed them entirely.
A guest file did not need to remember that someone “looked old.”
Priya also reviewed whether any guest note using age-related language had been copied into permanent guest profiles.
Most had not.
Where it had, Bellaforte replaced assumption-based labels with specific service preferences or removed them entirely.
A guest file did not need to remember that someone “looked old.”
Priya also reviewed whether any guest note using age-related language had been copied into permanent guest profiles.
Most had not.
Where it had, Bellaforte replaced assumption-based labels with specific service preferences or removed them entirely.
A guest file did not need to remember that someone “looked old.”
Priya also reviewed whether any guest note using age-related language had been copied into permanent guest profiles.
Most had not.
Where it had, Bellaforte replaced assumption-based labels with specific service preferences or removed them entirely.
A guest file did not need to remember that someone “looked old.”
Priya also reviewed whether any guest note using age-related language had been copied into permanent guest profiles.
Most had not.
Where it had, Bellaforte replaced assumption-based labels with specific service preferences or removed them entirely.
A guest file did not need to remember that someone “looked old.”
Priya also reviewed whether any guest note using age-related language had been copied into permanent guest profiles.
Most had not.
May you like
Where it had, Bellaforte replaced assumption-based labels with specific service preferences or removed them entirely.
A guest file did not need to remember that someone “looked old.”
