Tony Wang9 min readWhat Predicts a Google Rating Isn't Price. It's Whether You Book a Slot.
We measured ratings across 19 US industries and eight gym chains. Appointment-based businesses beat walk-in ones by half a star — and price barely matters.
The gym that costs 30x more and rates a star lower
Start with the case that makes the pattern visible, because gyms happen to price the same activity across a 30x range.
| Chain | Price/month | Format | Locations | Mean rating | vs. state baseline |
|---|---|---|---|---|---|
| Orangetheory | $160–200 | Booked class | 1,298 | 4.77 | +0.46 |
| Club Pilates | $150–200 | Booked class | 1,216 | 4.64 | +0.31 |
| Planet Fitness | $10–25 | Open gym | 2,614 | 4.34 | −0.06 |
| Anytime Fitness | $40 | Open gym | 2,526 | 4.33 | −0.10 |
| Crunch Fitness | $10–35 | Big-box | 581 | 4.15 | −0.23 |
| Gold's Gym | $30–70 | Big-box | 147 | 3.96 | −0.45 |
| LA Fitness | $40–70 | Big-box | 682 | 3.92 | −0.49 |
| Equinox | $300+ | Big-box | 119 | 3.76 | −0.61 |
Sort that table by price and it makes no sense: the $300 club is last, the $10 gym is mid-table, and the two $160 studios lead. Sort it by format and it is monotonic — booked classes, then open-access gyms, then big-box full-service clubs.
The mechanism this suggests is capacity. A booked class has a hard cap: you reserve one of twenty spots, so the experience cannot degrade by crowding. An open gym sells unlimited access to a fixed floor, and the peak-hour experience degrades no matter what the membership costs. Equinox charges thirty times more than Planet Fitness for a structurally identical promise — walk in whenever, use whatever is free.
The same pattern across 19 industries
If capacity and booking drive this, it should show up well beyond gyms. It does.
| Category | How you visit | Sampled | Mean rating | % below 4.0 |
|---|---|---|---|---|
| yoga_studio | Booked | 527 | 4.84 | 1.7% |
| day_spa | Booked | 285 | 4.73 | 3.9% |
| hair_salon | Booked | 561 | 4.72 | 3.7% |
| spa | Booked | 272 | 4.69 | 2.6% |
| gym | Mixed | 279 | 4.68 | 5.0% |
| self_storage_facility | Booked | 476 | 4.60 | 6.5% |
| auto_repair_shop | Booked | 280 | 4.57 | 8.2% |
| bakery | Walk-in | 259 | 4.49 | 7.7% |
| tire_shop | Mixed | 286 | 4.47 | 9.4% |
| ice_cream_shop | Walk-in | 295 | 4.45 | 9.8% |
| nail_salon | Mixed | 287 | 4.40 | 18.8% |
| coffee_shop | Walk-in | 582 | 4.39 | 17.0% |
| car_dealer | Mixed | 250 | 4.36 | 18.4% |
| cell_phone_store | Walk-in | 281 | 4.27 | 23.5% |
| urgent_care_center | Walk-in | 557 | 4.12 | 30.7% |
| pharmacy | Walk-in | 524 | 4.08 | 37.0% |
| hotel | Booked | 287 | 4.06 | 35.2% |
| bank | Walk-in | 561 | 3.79 | 53.7% |
| motel | Booked | 588 | 3.71 | 60.9% |
The four highest-rated categories are all appointment-based. The four lowest-rated where the visit mode is unambiguous — banks, pharmacies, urgent care, phone stores — are all walk-in-and-wait. Pooled, booked categories average 4.69 and walk-in categories 4.23.
Two exceptions are worth stating rather than hiding. Bakeries (4.49) and ice cream shops (4.45) are walk-in and rate well above the walk-in average — both are short, cheap, low-stakes visits where there is little to go wrong. And hotels (4.06) and motels (3.71) are booked and rate near the bottom, which is why they are marked separately below.
Why hotels are the control, not a counterexample
Hotels look like they break the rule until you notice they are the one category in this set where you are rating a product rather than an interaction. And there, price works perfectly.
| Tier | Brands measured | Locations | Mean rating | % below 4.0 |
|---|---|---|---|---|
| Luxury | Four Seasons, Ritz-Carlton, Waldorf Astoria | 224 | 4.48 | 7.1% |
| Upscale | Marriott | 2,702 | 4.22 | 12.5% |
| Midscale | Hampton Inn, Holiday Inn Express | 3,404 | 4.19 | 17.0% |
| Budget | Motel 6, Super 8, Days Inn | 4,035 | 3.44 | 86.7% |
A hotel room is capacity-managed by definition — you reserve a specific room for a specific night, so it cannot be oversold to you. What varies with price is the room itself, and the rating tracks it cleanly: 4.48 down to 3.44, with 86.7% of budget-motel locations below four stars. Restaurants behave the same way: Capital Grille 4.65 and Ruth's Chris 4.47 against Chili's 3.98 and Taco Bell 3.53.
So price predicts ratings fine when price buys a better thing. It fails in gyms because a more expensive membership does not buy a less crowded gym.
What we tested and discarded
Two hypotheses looked right and did not survive, which is worth recording because both are the obvious first guesses.
"Cheaper businesses rate higher." Dead on arrival once hotels and restaurants were measured — both are cleanly monotonic in the other direction. Gyms are the anomaly, not the rule.
"Contracts and lock-in breed resentment." This one felt strong: gyms are notorious for cancellation friction. But self-storage and telecom stores both lock customers into recurring contracts with price escalations, and they sit at opposite ends of the entire dataset — Extra Space Storage 4.85 and Public Storage 4.71, against Spectrum 3.03 and Xfinity 3.41. Identical contractual structure, 1.8 stars apart. Whatever is happening, lock-in is not it. (Both storage brands, notably, reserve you a specific unit; both telecom brands make you queue.)
How we did this, and the caveats
| Decision | What we did |
|---|---|
| Source | Crawlora's google-map-businesses dataset, 132M+ listings, US only (country=United States — the code US silently returns zero) |
| Ratings | The rating field, populated on ~97% of sampled records; unrated listings excluded |
| Brands | Counted by website domain, deduplicated by place_id; hotel brands sharing a parent domain matched on name and domain |
| Geography control | Each chain compared against the mean of other chains' locations in the same states, so no chain is scored against itself |
The visit-mode classification is ours, not the data's. Nothing in the dataset says whether a
business takes appointments. We classified the dominant mode by hand and marked the genuinely
mixed cases (nail_salon, tire_shop, car_dealer, gym) rather than forcing them. Reasonable
people would classify some differently, and the 0.46-star gap depends on where those lines fall.
Ratings measure who chooses to review, not satisfaction. Businesses that ask for a review at a happy moment will score higher than ones that don't, and we cannot separate that from genuine satisfaction. This is a plausible partial explanation for self-storage's 4.60 and it is not something this data can rule out.
Geography is controlled but not eliminated. State-level variation across gym chains runs 4.30 (Texas, New York, Washington) to 4.54 (Wisconsin, Alabama) — a 0.24-star spread against a 1.01 spread between chains, so the chain effect is roughly four times the state effect. Equinox is 37% New York and 29% California, whose gyms average 4.30 and 4.34; at 3.76 it sits well below its own markets.
Sampling was checked with a genuinely independent draw. Every category was sampled twice, from
result pages 1–3 and pages 40–42, with zero overlapping place_ids. All categories reproduced
within ±0.08 stars except motels (±0.24). Note that the dataset's sort parameter is silently
ignored when no q= term is supplied — identical order, identical set — so re-sorting is not a
valid way to draw a second sample here. We found that the hard way.
Price bands are approximate. Published US list prices vary by location, promotion and tier; treat them as bands, not quotes.
Measure a market yourself
The Google Business dataset behind this analysis carries rating, review count, category, website and geography on 132M+ listings — enough to test a hypothesis like this one in an afternoon. 2,000 free credits a month, no card.
Related reading
Frequently asked questions
Do appointment-based businesses really get better Google ratings?
In this sample, yes, and by a wide margin. Across 19 US categories, businesses where the dominant visit is booked average 4.69 stars while walk-in-and-wait businesses average 4.23 — a 0.46-star gap. The four highest-rated categories (yoga studios, day spas, hair salons, spas) are all appointment-based; banks, pharmacies, urgent care and phone stores sit at the bottom. The visit-mode classification is ours by hand, not a dataset field.
Does paying more get you a better-rated gym?
No — it inverts. Adjusted for the states each chain operates in, Equinox at $300+/month rates 0.61 stars below its local baseline and LA Fitness 0.49 below, while Orangetheory at roughly $160/month rates 0.46 above and Club Pilates 0.31 above. Planet Fitness at $10-25 sits mid-table at 0.06 below. Sorted by price the table is incoherent; sorted by format — booked class, open gym, big-box — it is monotonic.
Why do hotels follow price when gyms don't?
Because a hotel room is capacity-managed by definition: you reserve a specific room, so it cannot be oversold to you, and what price buys is a better room. Ratings track it cleanly — luxury 4.48, upscale 4.22, midscale 4.19, budget 3.44, with 86.7% of budget-motel locations below four stars. A gym membership is the opposite: it sells unlimited access to a fixed floor, so a more expensive membership does not buy a less crowded gym.
Isn't this just geography, since expensive gyms cluster in expensive cities?
Geography is real but far too small to explain it. Mean gym-chain ratings by state run from 4.30 (Texas, New York, Washington) to 4.54 (Wisconsin, Alabama), a 0.24-star spread, against a 1.01-star spread between chains. Every chain was scored against the mean of other chains' locations in the same states. Equinox is 37% New York and 29% California, whose gyms average 4.30 and 4.34; at 3.76 it sits well below its own markets.
Could contracts and cancellation friction explain the low gym ratings?
We tested that and it failed. Self-storage and telecom retail both lock customers into recurring contracts with price escalations, yet they sit at opposite ends of the whole dataset: Extra Space Storage 4.85 and Public Storage 4.71 against Spectrum 3.03 and Xfinity 3.41. Identical contractual structure, 1.8 stars apart. Notably both storage brands reserve you a specific unit while both telecom brands make you queue.