North America: the same list, three entirely different yardsticks

Three country pages, 59 rows, measured data for 53 accommodations. On the American "ten best" list the lowest review count is 102, on the Canadian one it is 5. The same page template, the same month, a twentyfold difference in what making the list means.

Beyond Europe · October 2026

In our article on Asia the main finding was that a "ten best" glamping list is often held up by a handful of reviews. In North America we measured the same page template: the country pages of the United States, Canada and Mexico, by machine, on 2026-10-09. In total 59 list rows, 53 different accommodations, each with name, type label, score and review count.

The result did not confirm the Asian finding. It took it apart. It turned out that the "handful of reviews" is not a property of the list type but varies by country, and between the three countries the difference is twentyfold.

The same two lists run on all three pages, but they do not mean the same thing

Two selections of ten stand one below the other: the "best 10 luxury campsites", and those "booked most often this month". On the "best 10" list the median review count:

On the American list not a single entry is below 100 reviews. On the Canadian one seven of the ten have 15 or fewer reviews, and among them three got their score from five:

This is what could not be seen from the Asian article. There, four country pages showed the same pattern, and from that it seemed that a small sample is the nature of the "ten best" selection. In North America it is the same page family, the same month, and the American list is made up of review counts in the two hundreds, the Canadian one of fives. In other words, the threshold for making the list does not come from the rule of the list, but from how many reviewed accommodations exist at all beneath the selection. Where there are few, five reviews are enough for the list of ten.

The score goes down as the sample grows, and this holds in all three countries

This is the second finding, and it is the one that does not change at the borders. Arranging the 53 unique accommodations into bands by review count:

Four bands, monotonically downward, a drop of 1.6 points. The co-movement of the two quantities across the 53 measured places is -0.655 (Pearson correlation between score and review count).

From this follows the gap between the two lists as well. On the "ten best" lists the median score of the 30 rows is 9.5, on the "most booked" lists the median of the 29 rows is 8.3. By country the difference is 1.30 (USA), 1.55 (Canada) and 0.40 (Mexico).

And there is a hard boundary: on the "ten best" list of all three countries every single entry is at 9.0 points or above, without exception. Meanwhile on the "most booked" list, in Canada ten of ten, in the United States six of ten, in Mexico five of nine entries are below 9.0.

So the "ten best" is not a ranking in the sense in which it is usually read. It is a score gate, through which places that have collected many reviews typically do not pass. A place that has served fifteen hundred guests will, statistically, not hold a 9.8. Not because it is worse, but because among fifteen hundred guests there are also those who were not satisfied.

The "luxury campsite" label is valid in New York too

The type label of all 59 rows is the same: "Luxury campsite" plus the name of the settlement. Not one other category occurs on the three pages. So the label says nothing about the kind of accommodation.

The sharpest example is the first place on the American "most booked" list: Carlton Arms Hotel, "Luxury campsite in New York", 7.0 points from 1785 reviews. This is the lowest score among the 53 measured places, and by its name it is a city hotel. In Mexico City, similarly, three different entries appear on the page: Mint Roma Residence, Casa Castillo Condesa, Roma Sur 222.

Counting the names: in 23 of the 59 rows a conventional accommodation word stands (Hotel, Inn, Motel, Auberge, Lodge, Cottages, Residence, Resort), and of these 16 are on the "most booked" lists, only 7 on the "best" lists. The word "glamping", on the other hand, appears in 11 rows in total, 9 of them on the "best" lists.

So the two lists are a different population not only in score but in accommodation type as well. By their names, the "best" selection tends to include genuine glamping places, the "most booked" list rather hotels, motels and inns. Whoever counts competitors from such a list is, on the second list, largely not counting glamping.

Where they are: the USA at park gates, Canada everywhere

The settlements reveal something about what the two markets offer. Read from the type label:

• United States: 20 rows, 14 different settlements. Six settlements occur twice, and four national parks stand in the names: Zion, Yosemite (twice), Grand Canyon (twice), Capitol Reef. Eight rows, six different places carry a park name.
• Canada: 20 rows, 20 different settlements. Not a single repetition, and the overlap between the two Canadian lists is zero entries.
• Mexico: 19 rows, 12 settlements, of which four rows fall on Mexico City.

In the USA the two lists overlap in 3 entries, in Mexico in 3, in Canada in none. The same page structure, three different outcomes.

What can be taken from this

1. The threshold of a "ten best" list is country dependent, not a property of the list. The same selection brings in a place with five reviews where the reviewed supply is scarce, and nothing below a review count in the hundreds where it is plentiful.
2. The high average and the high traffic are rarely the same place. Across the 53 measured places, score and review count move in opposite directions. Whoever plans for a 9.8 is not planning for high traffic.
3. The first reviews are worth disproportionately much, but not the same amount everywhere. Where supply is dense, you cannot even get into the selection below a hundred reviews.
4. Competitors cannot be counted from this label. Under "luxury campsite" there is a Manhattan hotel as well.

Sources, and what CANNOT be claimed from them

• Booking.com country glamping selections: `booking.com/glamping/country/us.html`, and the identically titled `ca` and `mx` pages. Retrieved 2026-10-09.
• From all three pages we collected the data by machine processing, not by reading: 20 rows from the American page and 20 from the Canadian one, 19 from the Mexican, in total 59 list rows and 53 unique accommodations. We recomputed the numbers from this saved collection, not from memory.
• The separation of the two lists was machine verified: the extraction confirmed the order of the two sections for all three countries, so the "ten best" and the "most booked" rows did not get mixed up.
• In the "most booked" section of the Mexican page there were nine entries, not ten. We note this here because the Mexican medians are computed from nine rows.

What these pages do NOT tell us: how many glamping accommodations there are in these countries, how big the market is, what the average price, the occupancy or the seasonality is. They are edited selections of ten items, not supply inventories, and not market size.

And what cannot be claimed about a row: a list row is not the same as an operator. One accommodation can appear in several rows, as in the USA and in Mexico three entries each stand on both lists at once, and behind one name there may be a chain, a single family or a city apartment. That is why we counted 59 rows and 53 unique places separately.

On the scores: the averages above 9 on these lists are not the averages of the market but those of the places brought into the selection, in Canada in several cases from five reviews. From this we infer neither a national standard nor a trend. The observation that the score moves downward with the review count is the internal pattern of these 53 places, from the moment of a single retrieval.