What 1,086 reviews say about the GTA sauna trade

Ratings cover 29 of the 39 companies in this index, and the 1,086 reviews behind them sit at a median of 12 per rated company. The mean rating across those 29 is 4.65 and the median is 5. Half the rated companies have 12 reviews or fewer, a sample small enough to move on a single review.

Of the 39 companies, 29 carry a rating, which is 74% of the roster. The other 10 show no reviews at all, because on a business listing the two arrive together: a rating is calculated from reviews, so a record with no reviews has no rating to show. Those 10 sit in the index with a name, a location and no score of any kind.

Those unrated records are not failing companies. An empty rating field means nobody has posted a public review of that business, which can happen to a new listing, to a business that sells through dealers rather than to the public, or to a trade whose customers do not think to review a supplier. Reading a blank as a bad sign inverts what the field says, and the index reports it as missing for that reason.

Set beside the other fields, the gap has a clear shape. The index holds a phone number for 38 of the 39 records, an address for 36 and a website for 30, and every record carries map coordinates. A profile shows the phone number only once the company claims its listing, and none has so far. Ratings, at 29 of 39, sit below all four of those fields, and no record carries an email address. The identity of these businesses is well documented. Public opinion about them is the thin part.

The total review count across the index is 1,086, and every one of them belongs to one of the 29 rated companies. That total is the whole body of public feedback behind the star column on this site, for every company it lists.

The median rated company holds 12 reviews, and 15 of the 29 hold 12 or fewer. Twelve reviews are enough to notice a pattern and not enough to trust one.

The maximum is 211 reviews on a single record, and the three most-reviewed companies hold 623 of the 1,086 between them, 57% of the total. Behind them sits a long run of records with a handful each: 6 of the 29 rated companies have 5 reviews or fewer. Averaging across the group hides that shape, which is why the median is the number worth carrying.

There is a second problem underneath the volume: the reviews do not all describe the same transaction. A review left after buying a heater off a showroom floor and a review left after a room was framed, wired and clad both land in the same average and carry the same weight. The index has no way to separate them, because it holds no job-completion data at all. A store with a strong counter and a builder with a strong crew look identical in this column.

Across the 29 rated companies in the index, the median rating is 5 and the mean is 4.65. Fifteen of the 29 sit at exactly 5, so the middle of the list is the top of the scale. There is no room above a 5, and the only direction any record can differ is down.

That is what pulls the mean below the median. The 14 scores below 5, two of them under 4, drag the average to 4.65 while the middle of the list stays pinned at 5. The full range runs from 2.3 to 5, and 23 of the 29 rated companies sit at 4.5 or above.

The consequence for a buyer is that ratings do little sorting here. With 15 of the 29 at 5, the score cannot separate them, and a gap of a fraction of a point between two records can come down to one review out of a dozen rather than a difference in how the two companies work.

This site's own published score does not take a star average at face value either. Its reviews pillar blends each rating with the 4.65 mean, as if every company had 10 extra reviews at that mean, then adds separate points for review volume on a logarithmic scale that tops out at 200 reviews. On that arithmetic a 5 on three reviews scores below a 4.8 on two hundred. The methodology page sets out the formula. The about page quotes a slightly different picture, an average of 4.6 and a median of 17 reviews across 26 companies, because it counts only companies with three or more Google reviews.

The low end of the range is 2.3, and it belongs to a single company with 3 reviews. That figure carries the same warning as every other score in the set: a low average on a small sample is as unstable as a high one, and it moves as far on one review in either direction.

The index publishes it without comment because hiding the low end would misrepresent the range. What it does not do is treat it as a verdict. This site has no job-completion data, no complaint records and no way of knowing what produced any particular score, and calling one company the weakest in a market on the strength of three reviews would be a claim the data cannot support.

A blank and a low score deserve different treatment, and buyers often give them the same one. A low average is information, thin but real. A blank is the absence of information, and a business that sells through dealers or works by referral can sit at zero public reviews. Neither should end a shortlist on its own.

The fair use of the bottom of the range is as a scale marker. It tells a reader that scores in this index reach as low as 2.3, so the 4.65 mean across the 29 rated companies sits above a real spread rather than a floor.

The purchases these ratings sit next to are large. RenoHouse (Toronto) prices a prefab Finnish sauna, professionally installed, all in, at $10,000 to $20,000, and a semi-custom Finnish sauna with full basement integration at $18,000 to $32,000. The small end of the market is not trivial either: Relaxacare, a Burlington, ON retailer, lists a 1 to 2 person Westinghouse far-infrared cabin at $2,499.99.

A star average built on 12 reviews is being asked to underwrite decisions at those amounts, and it cannot. What it can do is tell a buyer that a business exists, has customers who post, and has not collected an obvious pattern of complaints. That makes it a screening tool rather than a selection tool, and it belongs at the start of a shortlist rather than at the end of one.

The fields that would settle a purchase are the ones the index does not hold. There is no job-completion data here, no timelines, no warranty terms, no licence status, no revenue, no employee count and no years in business for any company. None of it is in the records this index is built from, and inventing it would produce a page that looks more useful and is worth less.

Treat the review count as the first number and the star average as the second. A 4.6 across a large body of reviews and a 5 across three are not comparable readings, and the index prints the two figures side by side so the comparison can be refused rather than made by accident. A record with no rating deserves a look at the company's own website rather than an assumption.

The second habit worth having is checking when the reviews arrived, which the index does not record but the business listing itself shows. A count of 12 collected slowly over years and a count of 12 collected in a single burst describe different businesses, and the average reported here cannot tell the two apart.

Then check the score against something outside it. What a company sells, and whether its own website describes installation work or equipment sales, are more decisive than a fraction of a star. Of the 39 companies in the index, 38 sit in Saunas and 22 in Cold Plunge & Ice Baths, and the two counts overlap because a company can hold both. The Toronto page, with the six companies filed under Willowdale, and Vaughan, with 4, are where a buyer can compare several records at once. The methodology page explains how the site's own ranking handles thin review counts.