How the score works, in full
A number nobody can check is worth nothing. Here is the whole formula, the reasoning behind the one part of it that is a judgement call, and a worked example you can redo by hand.
Two kinds of number, never mixed
Measured means the merchant published it and we read it: the stated lift, the price, the sizes, whether it is in stock. We convert units and we do arithmetic on it, and that is all.
Modelled means we decided it. There is exactly one modelled input on this site and it is described below. Everywhere it appears, it is labelled.
We have not worn these shoes. Nobody who has worn 1,590 pairs of elevator shoes is writing honestly about all of them. So there is no comfort rating, no durability rating and no star review of our own. Where a merchant publishes its own customer rating we show it and say whose it is.
The formula
Every shoe starts at zero and collects points from four places. The maximum is 100.
1. Value for the lift, up to 40 points
Price divided by centimetres of lift gives a cost per centimetre. That figure is ranked against every other shoe in the database, and the points are the percentile: cheapest per centimetre scores 40, most expensive scores 0.
This is deliberately the heaviest part. It is the one comparison the category never makes about itself.
2. Whether it stays believable, up to 25 points
This is the modelled part. A shoe can only hide so much sole before it stops looking like a shoe, and how much depends entirely on its shape. We give each shape a figure for how much lift it can carry:
| Shape | Carries up to | Why |
|---|---|---|
| Boots | 9 cm | A shaft covers the ankle, so a deep sole has somewhere to go |
| Sneakers | 8 cm | A thick midsole is normal on a trainer, so nobody reads it as a lift |
| Casual shoes | 7 cm | Some sole depth is expected, but not much |
| Dress shoes | 6.5 cm | A formal last is thin by convention, so the lift has to sit inside the shoe |
| Insoles | 4 cm | Limited by the room left inside a shoe built for a foot |
| Sandals | 2.5 cm | An open shoe hides nothing at all |
Points are 25 × min(1, carries ÷ lift). A boot with 7 cm of lift takes the full 25 because 9 is more than 7. A dress shoe with 13 cm takes 13.
This is a judgement, and we would rather show you the judgement than bury it in a star rating.
3. Whether you can actually buy it, up to 20 points
In stock takes 20. Low stock takes 9. Sold out takes 0. Where a merchant does not publish stock per size, the shoe is capped at 12 rather than given the benefit of the doubt.
4. Whether it fits more feet, up to 15 points
The number of sizes listed, against a reference of eight. Where sizes are not published in a form we can read, the shoe is capped at 9 rather than guessed at.
Those two caps mean a shoe from a merchant with thinner data cannot reach 100. That is the correct outcome. Less information should not score the same as good information.
A worked example
Take Q212 (Black/Grey) Ultra Lightweight, at $37.49 with a stated 6.1 cm lift.
| Part | Working | Points |
|---|---|---|
| Value for the lift | $6.15 per cm of lift. Cheaper than 97% of the 1,588 shoes we hold. | 39 / 40 |
| Stays believable | 6.1 cm of lift in a sneaker style, which we rate as able to carry 8 cm before it reads as a lift. | 25 / 25 |
| Actually buyable | 7 of 8 sizes in stock at the last check. | 20 / 20 |
| Fits more feet | 8 US sizes listed. | 15 / 15 |
| Total | 99 / 100 |
Where the data comes from
Two merchants publish enough structured detail to compare honestly: Chamaripa and TallMenShoes.com. We read their own product data, we do not retype it, and we rebuild the whole database rather than editing rows by hand. The date at the foot of every page is when that last happened.
Prices move. Stock moves faster. If a figure here disagrees with the merchant's own page, the merchant is right and we are stale.
How we make money, plainly
Some outbound links will eventually carry an affiliate tag, which pays us a commission at no cost to you. Right now most of them are plain links that pay us nothing. Either way the ranking is produced by the formula above, which has no input for what a merchant pays. The full disclosure is here.