Marriott.
Guest Experience Intelligence · powered by Dimension Labs
Live · refreshes weekly
Causal Intelligence · Orlando Portfolio

Guest Experience Causal Monitor

45 hotels · trailing 4 weeks to July 1, 2026 · 512 new reviews scored. Every metric is ranked by what it causes, not by how often it is mentioned, and every move is tested for significance before it earns an arrow.
Tested: 1 of 7 period moves is real · resort-fee complaints are climbing
This period at a glance
Read this first

A normal guest-experience board ranks problems by how often guests mention them, and sends you to the front desk, the most-mentioned topic at 27% of comments. The causal layer says the front desk moves the outcome by almost nothing (+2pt, not significant). The real drivers of a poor stay are cleanliness (+34pt), staff conduct (+44pt), and billing disputes (+19 to +44pt). This monitor tracks the causes, flags which moves are real, and shows what closing the largest gap is worth. Tap any card to expand it.

What actually moved this period — one thing did.

Each metric's change vs the prior 4-week window, with its 95% confidence interval. Six moves stay inside the noise band; resort-fee complaints clear it, the one detectable change this week. Two-proportion z-tests · n=512 reviews.
≈ SAMPLING NOISE0 — no change−10pp+10ppPoor-stay raten.s.Dirty-room signaln.s.Staff-conduct signaln.s.Billing-dispute signaln.s.Front-desk mentionsn.s.Resort-fee complaintsREAL ↑Named-hero raten.s.Six of seven moves are noise. One clears its band: resort-fee complaints, at +3.9pp.
real · clears its band within noise · n.s.α=0.05

Alerts, reframed

Don't chase the front-desk mention rate.
The loudest topic on any board, but its causal effect on a poor stay is +2pt and not significant. Staffing an alarm to it spends attention on a mirage.
→ Rank by causal effect, not by volume.
Cleanliness is the standing #1 driver.
A dirty-room signal raises poor-stay risk +34pt (CI 30.6–37.4), stable for 18 months. The largest single lever on the public score.
→ Manage it where the sample is firm.
A wrong charge is the most damaging event.
+44pt per mention, but on a thin base (n=98). Directionally severe, confirm as volume grows before setting a target.
→ Watch the billing-dispute signal.
Causal driver scorecard — what is causing the score, tracked weekly
A dirty room makes a one-to-two-star stay about three times more likely (+34 points). Cleanliness and staff conduct are the two standing priorities, the front desk is a mirage, and a named hero is the growth lever. The board below tracks each cause every week.
DriverShare of stays8-week trendCausal effect on a poor stayConfidenceStatus
Cleanliness
A dirty or unclean room
+34pt
StrongPriority
Staff conduct
Rude or unprofessional
+44pt
StrongPriority
Something broken
In-room maintenance
+21pt
ModerateMonitor
Surprise fee
An undisclosed resort fee
+19pt
ModerateRising
Wrong charge
A disputed folio charge
+44pt
Moderate · thin nWatch
Reservation not honored
A booking not kept
+29pt
Moderate · thin nWatch
Front desk / check-in
Most mentioned, no effect
+2pt
No effectMirage
Named hero
An employee who went beyond
−17pt
ModerateProtect
Proof of rigor — is the driver real, or a coincidence?

Shuffle the cause and it vanishes. That is how we know.

Every effect is stress-tested three ways. Pick a driver.

Placebo = the treatment label is randomly shuffled. A real cause collapses to zero here. Random-cause and subset re-run the model with an injected noise variable and on a random half of the data; a real effect barely moves.

Rank by cause, not by volume

The loudest topic causes the least.

Toggle the ranking. The order of the bars flips, front desk drops from first to last.
Counterfactual — what closing the gap buys

Drag to size the prize before you spend a dollar.

Close the dirty-room gap at the five watch hotels toward the portfolio median and watch the projected poor-stay share fall. Computed live from the +34pt causal effect and the real per-hotel review volumes.
Now
With fix
poor-stay share prevented
poor stays prevented per period (512 reviews)
room-rate headroom, illustrative (Cornell)
projected portfolio poor-stay share at the selected fix level. Treat as the center of a range, not a promise.
The Meaning Layer — one review becomes structured proof
★★★★
We booked a king suite for our anniversary and arrived to a room that had not been touched: hair in the sink and a stained carpet. When I asked to move, the manager was dismissive and made us feel like we were the problem. Then a $39 resort fee we were never told about. We will not be back.
Guest review · Courtyard LBV / Marriott Village · Jun 2026 · one raw paragraph → nine structured fields
23 dimensions extracted from this one review
Overall sentiment
Strongly negative
Primary friction
Cleanliness
Cleanliness signal
Broadly dirty
Staff interaction
Named blame
Staff role
Front-desk manager
Fee signal
Surprise resort fee · $39
Trust signal
Alleged non-disclosure
Loyalty signal
Leaving the brand
Recommendation
Will not return
Hover a highlighted phrase to see the field it became. A survey gives you a star and a checkbox; the Meaning Layer turns the sentence into queryable structure, which is what makes the causal read possible.
Voice of the guest — the evidence behind every driver
Why this matters

A number tells you the size of a cause; a verbatim tells you the mechanism, in the guest's own words. Every quote below was tagged by the Meaning Layer and routed to the driver it evidences. Filter by driver to read the proof, and notice that front-desk mentions show up on five-star reviews too, which is exactly why they cause nothing.

★★★★
Beautiful lobby, but the room told a different story. Mold along the tub, a stained carpet, and hair on the sheets. For a Sheraton I expected far better.
Sheraton Orlando Lake Buena Vista · Jun 2026
Sentiment: Strongly negPrimary friction: CleanlinessCleanliness: Broadly dirtyRecommend: Will not return
Meaning Layer → evidences: Cleanliness · +34pt driver
★★★★★
Housekeeping skipped our room two days running. The bathroom was never restocked and the floor was visibly dirty. Front desk apologized but nothing changed.
Delta Hotels Orlando Celebration · Jun 2026
Sentiment: NegativeCleanliness: Minor lapseIssue resolution: Not fixed
Meaning Layer → evidences: Cleanliness · +34pt driver
★★★★
Room smelled musty the second we opened the door and the AC vent was packed with dust. We asked to switch and the next room was no better.
Courtyard LBV / Marriott Village · May 2026
Sentiment: Strongly negCleanliness: OdorRoom condition: Worn
Meaning Layer → evidences: Cleanliness · +34pt driver
★★★★
The front-desk manager rolled her eyes when I asked about a charge and told me to read the confirmation. Made us feel like a nuisance on our anniversary.
Sheraton Suites Orlando Airport · Jun 2026
Sentiment: Strongly negStaff interaction: Named blameStaff role: ManagerRecommend: Warn others
Meaning Layer → evidences: Staff conduct · +44pt driver
★★★★★
Called the desk three times about noise next door. The overnight staff was short and dismissive every time and never sent anyone up.
SpringHill Suites Orlando Airport · May 2026
Sentiment: NegativeStaff interaction: Unnamed blameIssue resolution: Escalated
Meaning Layer → evidences: Staff conduct · +44pt driver
★★★★★
Hit with a $39/night resort fee that was never disclosed at booking. No resort, no pool worth the name, just a fee. Felt like a bait and switch.
Delta Hotels Orlando Celebration · Jun 2026
Sentiment: NegativeFee signal: Surprise resort feeTrust: Alleged non-disclosure
Meaning Layer → evidences: Surprise fee · +19pt driver
★★★★
They double-charged my card and it took four calls and two weeks to reverse. Nobody could explain it. I will dispute future stays before I trust the folio again.
Sheraton Orlando Lake Buena Vista · Apr 2026
Sentiment: Strongly negFee signal: Billing errorTrust: BrokenLoyalty: Leaving brand
Meaning Layer → evidences: Wrong charge · +44pt driver
★★★★
Booked a king suite months ahead. Arrived and they had no record and put us in a double by the elevator. Elite status meant nothing that night.
Courtyard LBV / Marriott Village · May 2026
Sentiment: Strongly negReservation: Not honoredLoyalty: Elite ignored
Meaning Layer → evidences: Reservation broken · +29pt driver
★★★★★
The AC rattled all night and the toilet ran constantly. Maintenance was called twice and never showed. Exhausted before the parks even opened.
SpringHill Suites Orlando Airport · Jun 2026
Sentiment: NegativeRoom condition: BrokenIssue resolution: Not fixed
Meaning Layer → evidences: Something broken · +21pt driver
★★★★★
Check-in line was a little long at 4pm, but the front desk sorted us a room upgrade and the rest of the stay was flawless.
The Ritz-Carlton Orlando, Grande Lakes · Jun 2026
Sentiment: PositiveFront desk: MentionedOutcome: 5-star
Meaning Layer → evidences: Front desk · +2pt (the mirage)
★★★★
Front desk was busy at check-in during a convention, though they were friendly and the stay itself was great.
Orlando World Center Marriott · May 2026
Sentiment: PositiveFront desk: MentionedOutcome: 4-star
Meaning Layer → evidences: Front desk · +2pt (the mirage)
★★★★★
Front desk check-in was slow, but honestly the real problem was the dirty room and the surprise fee.
Delta Hotels Orlando Celebration · Jun 2026
Sentiment: NegativeFront desk: MentionedPrimary friction: Cleanliness
Meaning Layer → evidences: Front desk · +2pt (the mirage)
★★★★★
Maria at the front desk surprised our daughter with a birthday treat and balloons. We are already booking next year because of her.
JW Marriott Bonnet Creek Resort · Jun 2026
Sentiment: Strongly posStaff named: MariaPraise driver: A personRecommend: Loyal repeat
Meaning Layer → evidences: Named hero · +42pt five-star lever
★★★★★
Our breakfast server James remembered my parents' order every morning and made them feel like royalty. That is why we come back to this Ritz.
The Ritz-Carlton Orlando, Grande Lakes · May 2026
Sentiment: Strongly posStaff named: JamesPraise driver: A personLoyalty: Repeat
Meaning Layer → evidences: Named hero · +42pt five-star lever
★★★★★
Diego from the concierge team booked a last-minute dinner and arranged a car when ours fell through. An absolute professional who saved our night.
Grand Bohemian Orlando, Autograph · Jun 2026
Sentiment: Strongly posStaff named: DiegoStaff role: ConciergePraise driver: A person
Meaning Layer → evidences: Named hero · +42pt five-star lever
★★★★★
Housekeeper Rosa left a handwritten note and towel animals for the kids. Small thing, huge impact. This is why this is our home base in Orlando.
TownePlace Suites Orlando Downtown · Apr 2026
Sentiment: Strongly posStaff named: RosaStaff role: HousekeepingRecommend: Loyal repeat
Meaning Layer → evidences: Named hero · +42pt five-star lever
★★★★★
Room was immaculate, the bed was heaven, and the pool spotless. Everything worked, nothing surprised us on the bill. Best Marriott stay we have had.
JW Marriott Bonnet Creek Resort · May 2026
Sentiment: Strongly posCleanliness: SpotlessPraise driver: Cleanliness
Meaning Layer → evidences: Delight · five-star driver
★★★★★
Spotless, quiet, and the staff anticipated everything. Not one thing went wrong. We will absolutely return.
The Westin Lake Mary, Orlando North · Jun 2026
Sentiment: Strongly posCleanliness: SpotlessRecommend: Loyal repeat
Meaning Layer → evidences: Delight · five-star driver
The named hero — proof of the growth lever

The growth lever, in guests' own words

A guest naming an employee who went beyond is the single strongest cause of a five-star review (+42pt) and it lowers poor-stay risk by 17 points. It is the cheapest five-star you can manufacture, and it is invisible to a survey that never asks for a name.
Maria · Front desk
★★★★★
"Surprised our daughter with a birthday treat and balloons. We are already booking next year because of her."
JW Marriott Bonnet Creek
James · Dining
★★★★★
"Remembered my parents' order every morning and made them feel like royalty."
The Ritz-Carlton, Grande Lakes
Rosa · Housekeeping
★★★★★
"Left a handwritten note and towel animals for the kids. Small thing, huge impact."
TownePlace Suites Downtown
Where the mechanism fires — all 45 hotels
How to read

The gap is not evenly spread. Sort by poor-stay rate to find where to act first, filter by tier, or search a property. A high rate on a large review base is a firm finding; the best-in-class rates sit on thinner samples, confirm before you benchmark to them.

Hotel Tier Reviews Avg star Poor-stay rate Flag
Recommended actions this period
Action 01

Fix cleanliness where it's firm.

Send a housekeeping audit to the five watch hotels. Cleanliness is the largest causal lever on the public score (+34pt), and the gap is concentrated, not portfolio-wide.

Moves: dirty-room rate → poor-stay share
Action 02

Stop staffing the front-desk alarm.

Front desk is 27% of mentions and near-zero causal effect. Rank the board by causal effect so the team stops chasing the loudest signal instead of the strongest one.

Moves: attention, alert precision
Action 03

Protect and scale the named hero.

A guest naming an employee who went beyond both prevents a one-star (−17pt) and drives a five-star (+42pt). Make it a coached, recognized behavior.

Moves: named-hero rate → five-star share

On the watchlist — signal, not noise

  • Cleanliness +34pt (CI 30.6–37.4): the causal parameter to protect. If the dirty-room rate climbs, the public score follows.
  • Wrong charge +44pt but only n=98, a thin base. The most damaging event, confirm as volume grows before setting a target.
  • The front-desk mirage: keep confirming it stays near zero. Until it clears its band, don't fund it.
  • The named-hero rate: the one growth lever. Watch it rise where you coach it, the cheapest five-star you can make.
Marriott Guest Experience Causal Monitor · built on Dimension Labs Causal Intelligence · DoWhy ATE + propensity-score weighting, two-proportion z-tests, placebo / random-cause / subset refutation · α=0.05 · n=8,504 scored reviewsTemplate · wire to your live review feed