A survey and a review dashboard rank issues by volume. This brief ranks them by cause, for the three KPIs a hotel owner actually manages. The order is not what your inbox would predict.
Ranked by volume vs. ranked by cause
Left: the order a survey or review tool shows you, by how often each issue is mentioned. Right: the same issues ranked by their measured causal effect on a one- or two-star review. The order nearly inverts.
Mention share vs DoWhy causal effect, 8,504 rated reviews. Front desk falls from first to last.
This is the entire argument of the brief, in one exhibit. Everything that follows applies the same causal engine to the three outcomes a hotel owner actually manages: the reviews that sink a score, the reviews that build one, and the guests who quietly never come back.
The causal drivers of a 1–2 star review
Each bar is the measured effect of a signal on the probability a stay ends in one or two stars, isolated from the hotel, tier, language, review length, and reviewer.
DoWhy causal model, 8,504 rated reviews. Bar = point estimate, whisker = 95% confidence interval, right column = confidence grade. All effects refutation-tested.
In the guests’ own words, the causes are unmistakable, and they are operational, not cosmetic:
“My daughter is 16 years old and is traveling from Miami with her school to a HOSA event. It's 7:00 pm and they haven't given her her room. She had to get ready in the public bathroom. This hotel leaves a lot to be desired. Don't even think about coming here.”★ · Orlando World Center Marriott
“Horrible room service. They don't stock the bathroom with enough toilet paper. They don't replenish the coffee refills, etc. I do not recommend it.”★ · Delta Hotels Orlando Celebration
“The worst hotel ever, and I don't want to hear from them. We regret what happened because everything was paid for, and on top of that, they kept the deposit. It's all a rip-off with the card, from the moment they answer the phone to the manager. The worst.”★ · Delta Hotels Orlando Celebration
“The hotel is terrible, the staff service is the worst, they are rude and never want to help. Beware of staying at this hotel.”★ · Residence Inn Orlando Airport
A team working the complaint count would add lobby staff and never touch housekeeping or the folio, the two things actually manufacturing the one-star tail.
The causal drivers of a 5-star review
The same engine, run on the winning outcome. A human being is the top driver; a spotless room second; price and location at the bottom.
DoWhy causal model on the five-star outcome, 8,504 rated reviews. The strongest lever, an individual person, is the one no survey structure isolates.
And guests name the people who own their best stays, which is exactly the signal a survey throws away:
“Everything was great! The hotel is really remote, it's very idyllic, and the shuttle bus to the parks is amazing!.”★★★★★ · The Ritz-Carlton Orlando
“Wonderful hotel, the room, the breakfast, the pool, everything was wonderful, as well as the attention of the chambermaid Claudia Amaya was magnificent.”★★★★★ · JW Marriott Orlando Bonnet Creek Resort & Spa
“The place is very beautiful, one of the most privileged spots in Orlando, excellent location and built with the finest quality details.”★★★★★ · The Ritz-Carlton Orlando
“Fantastic hotel. Great pool area. Excellent service. Comfortable bed. Good food, including breakfast. Would definitely stay again. Thumbs up!.”★★★★★ · JW Marriott Orlando
Cleanliness and maintenance set the floor of your score; a nameable person sets the ceiling. Make that service a standard and you manufacture five-star reviews on purpose.
The causal drivers of ‘I will not return’
The effect of each signal on the probability a review carries an explicit will-not-return or warn-others signal. Trust breaches dominate; a one-off defect is far lower.
Causal model on the will-not-return signal from the Meaning Layer, 8,504 rated reviews. This is the retention outcome a star score cannot see directly.
The language of a lost guest is not about a bad night. It is about being wronged:
“The hotel is fine, just be careful if you make and pay for your reservation through a platform like Hotels.com or Expedia because the hotel doesn't honor this and will charge you again for the room, even if you explain the situation, they won't refund anything.”★ · The Ritz-Carlton Orlando
“Beware of these hotels; don't fall for their park discount scam. They make you attend a 3-hour talk, wrap you up with packages, and then they never show up.”★ · Courtyard by Marriott Orlando LBV
“They charge $42 per day for parking, and in addition to the room rate, there's an extra charge they don't tell you about at check-in: a resort fee of $45 per person per day. Watch out! Those fees aren't listed in the Marriott Bonvoy app. Very disappointed.”★ · Renaissance Orlando at SeaWorld
The pattern is consistent and it is a warning: a dirty room makes a guest unhappy, but a charge they did not authorize makes them done. Protect the folio and the reservation with the same rigor you protect the room.
The full portfolio, worst to best, adjusted for sample size
Each hotel’s one- and two-star rate, empirical-Bayes shrunk so small hotels are not over-punished. The top is the watch list; the bottom is the proof of what good looks like, same brand, same city.
| Hotel | Tier | Reviews | Avg star | Raw 1–2★ | Adjusted | |
|---|---|---|---|---|---|---|
| Sheraton Orlando Lake Buena Vista Resort | Full-Service | 293 | 2.92 | 47.4% | 45.9% | watch |
| Courtyard by Marriott Orlando LBV in the Marriott Village | Select-Service | 168 | 2.90 | 46.4% | 44.0% | watch |
| Sheraton Suites Orlando Airport | Full-Service | 140 | 3.17 | 41.4% | 39.1% | watch |
| Delta Hotels Orlando Celebration | Select-Service | 540 | 3.31 | 35.4% | 34.9% | watch |
| Residence Inn Orlando at SeaWorld | Extended-Stay | 180 | 3.35 | 35.0% | 33.8% | watch |
| Fairfield Inn Orlando Lake Buena Vista in the Marriott Village | Budget | 212 | 3.28 | 34.4% | 33.5% | |
| SpringHill Suites Orlando Airport | Select-Service | 116 | 3.29 | 32.8% | 31.3% | |
| Sheraton Orlando North Hotel | Full-Service | 228 | 3.50 | 29.8% | 29.3% | |
| Castle Hotel, Autograph Collection | Luxury | 226 | 3.65 | 28.3% | 27.9% | |
| Renaissance Orlando Airport Hotel | Full-Service | 138 | 3.53 | 28.3% | 27.6% | |
| Fairfield Inn Orlando at SeaWorld | Budget | 138 | 3.57 | 28.3% | 27.6% | |
| Fairfield Inn Orlando International Drive/Convention Center | Budget | 120 | 3.62 | 26.7% | 26.2% | |
| Fairfield Inn Orlando Airport | Budget | 170 | 3.58 | 26.5% | 26.1% | |
| SpringHill Suites Orlando at Millenia | Select-Service | 82 | 3.78 | 26.8% | 26.1% | |
| Four Points by Sheraton Orlando International Drive | Select-Service | 274 | 3.58 | 25.9% | 25.7% | |
| Residence Inn Orlando Airport | Extended-Stay | 84 | 3.63 | 26.2% | 25.6% | |
| Fairfield Inn Orlando Kissimmee/Celebration | Budget | 93 | 3.61 | 25.8% | 25.3% | |
| Renaissance Orlando at SeaWorld | Full-Service | 371 | 3.70 | 25.3% | 25.2% | |
| Residence Inn Orlando Lake Buena Vista | Extended-Stay | 213 | 3.85 | 24.4% | 24.3% | |
| Residence Inn Orlando Convention Center | Extended-Stay | 64 | 3.80 | 23.4% | 23.4% | |
| Courtyard by Marriott Orlando International Drive/Convention Center | Select-Service | 80 | 3.73 | 22.5% | 22.6% | |
| Marriott Orlando Downtown | Full-Service | 175 | 3.86 | 22.3% | 22.4% | |
| SpringHill Suites Orlando at Flamingo Crossings | Select-Service | 196 | 3.89 | 21.4% | 21.6% | |
| SpringHill Suites Orlando Convention Center/International Drive Area | Select-Service | 117 | 3.90 | 20.5% | 20.9% | |
| Orlando World Center Marriott | Full-Service | 643 | 3.95 | 20.1% | 20.2% | |
| Element Orlando Universal Blvd | Extended-Stay | 158 | 4.09 | 19.0% | 19.5% | |
| Grand Bohemian Orlando, Autograph Collection | Luxury | 264 | 4.07 | 18.9% | 19.2% | |
| TownePlace Suites Orlando at Flamingo Crossings | Extended-Stay | 153 | 3.99 | 17.6% | 18.3% | |
| Courtyard by Marriott Orlando South/Grande Lakes Area | Select-Service | 79 | 3.95 | 16.5% | 17.8% | |
| Marriott Orlando Airport Lakeside | Full-Service | 282 | 4.18 | 16.3% | 16.8% | |
| AC Hotel Orlando Downtown | Select-Service | 151 | 4.09 | 15.9% | 16.7% | |
| Fairfield Inn Orlando at Millenia | Budget | 95 | 4.15 | 14.7% | 16.2% | |
| The Westin Lake Mary, Orlando North | Full-Service | 216 | 4.31 | 15.3% | 15.9% | |
| Aloft Orlando Downtown | Select-Service | 155 | 4.30 | 14.8% | 15.8% | |
| Courtyard by Marriott Orlando Altamonte Springs/Maitland | Select-Service | 54 | 4.17 | 13.0% | 15.7% | |
| Courtyard by Marriott Orlando Airport | Select-Service | 95 | 4.05 | 13.7% | 15.3% | |
| Residence Inn Orlando Downtown | Extended-Stay | 82 | 4.26 | 12.2% | 14.3% | |
| TownePlace Suites Orlando Airport | Extended-Stay | 110 | 4.36 | 12.7% | 14.3% | |
| JW Marriott Orlando, Grande Lakes | Luxury | 366 | 4.22 | 13.7% | 14.2% | |
| Courtyard by Marriott Orlando Downtown | Select-Service | 126 | 4.29 | 12.7% | 14.1% | |
| The Ritz-Carlton Orlando, Grande Lakes | Luxury | 325 | 4.33 | 13.5% | 14.1% | |
| TownePlace Suites Orlando Downtown | Extended-Stay | 64 | 4.47 | 9.4% | 12.7% | |
| Residence Inn Orlando at Flamingo Crossings Town Center | Extended-Stay | 161 | 4.42 | 11.2% | 12.5% | proof |
| Fairfield Inn Orlando at Flamingo Crossings Town Center | Budget | 192 | 4.40 | 10.9% | 12.1% | proof |
| JW Marriott Orlando Bonnet Creek Resort & Spa | Luxury | 315 | 4.53 | 9.2% | 10.0% | proof |
Empirical-Bayes shrinkage across 45 hotels. Raw = observed share, Adjusted = shrunk estimate.
The two worst hotels sit in the same submarket as mid-pack sisters that run clean. The problem is not Orlando and it is not the tier. It is the property, which means the fixes in this brief are enforceable one hotel at a time.
Cleanliness (+34) and staff conduct (+44) drive the one-star tail; the front desk (+2) does not. Re-point audits and staffing at the two real causes.
A wrong charge or an unkept reservation is the top cause of a guest never returning, far above any defect. Protect the folio and the booking with the same rigor as the room.
A named employee is the one lever that both prevents a one-star and manufactures a five-star. Recognize it, coach it, reproduce it across the middle tiers.
A survey gives you a star and a checkbox. Dimension Labs reads each review into 23 structured signals, from a named employee to a withheld deposit, which is what makes every causal driver above possible. This is the dictionary.
Every figure is derived from 9,446 public Google reviews of 45 Marriott-family hotels in the Orlando metro, January 2025 to July 2026, deduplicated to one row per review. No survey or proprietary data was used.
The Meaning Layer. Each review was enriched into the 23 structured signals above, turning unstructured text into queryable data.
The causal engine. Driver effects for each KPI come from a DoWhy average-treatment-effect model on the 8,504 rated reviews, with backdoor adjustment for hotel, tier, language, review length, and reviewer, cross-checked with propensity-score weighting and stress-tested with placebo, random-common-cause, and subset refutation. The never-return board is modeled on the will-not-return signal and is a directional causal read.
The portfolio read. The local-versus-national split is a mixed-effects variance decomposition; the hotel ranking is empirical-Bayes shrunk. Public company facts (Marriott FY2025 results, the FTC fee rule) are cited to their sources and never merged with the review estimates.