Methodology

What is rated

A passenger rates one landing on a scale of 0.5 to 5 stars in half-star steps. 5 is the smoothest. They may add a description (smooth, firm, hard, bounced or go-around), whether passengers applauded, and whether there was turbulence on approach. The form only offers descriptions that fit the star rating, so a 4.5-star landing cannot be tagged as bounced.

The Touchdown Quality Index (TQI) for one rating is the star rating divided by 5 and multiplied by 100. A 4-star landing is a TQI of 80. Nothing else is added to it.

Ratings fall into three groups by star value: smooth (4 and above), firm (2.5 to 3.5) and rough (below 2.5).

What is not measured

The index records what passengers perceived. It does not record touchdown vertical speed or g-load. Public ADS-B tracking updates every few seconds and rounds altitude to 25 ft, which is too coarse to see how firm a touchdown was. Seat position, aircraft type, wind and passenger expectations all affect what a passenger reports, and the index does not correct for them.

If measured approach data is added later, it will appear as a separate figure with its own method. It will not be blended into the passenger score.

The signed-in label

Each rating carries one of two labels. Signed in means the person was signed in with Google when they submitted. Guest means they submitted without an account; the rating is filed under a random guest code stored in a cookie on that browser, so ratings from one browser can be grouped.

The label does not confirm that the flight took place. There is no booking or boarding-pass check. Signed-in users can rate a given carrier, route and date once; a guest code identifies a browser and not a person, so a guest can rate a given carrier, route and date once per browser, and anyone who clears cookies can rate again. Each browser or account is also limited to 40 ratings in 24 hours. Both kinds count toward scores, and every page shows how many of its ratings are signed in.

A rating can also show flight matched. That means the flight number and date were looked up in an airline schedule source and the route returned (origin and destination) is the one the person entered. It is a check that the flight existed on that route and date, not that the person was on it. Many ratings will not carry the mark, because a match is only recorded when the lookup succeeds.

Reviews policy

Ratings are submitted through the form on this site, signed in with Google or as a guest. One rating covers one landing. Nobody is paid, given a discount, entered into a prize draw or otherwise rewarded for rating, and nothing is offered for rating a carrier or airport up or down. Airlines and airports are not asked to rate or to encourage rating.

Every submission is checked against the same limits: 12 ratings per 10 minutes from one network address (or 50 rows from an imported flight log), 40 in 24 hours per account or guest code, and one rating per carrier, route and date per account or browser. A rating with invalid details is rejected, as described under corrections.

Ratings can also be added from a flight log on the import page. These are scores for past flights, given from memory, and only flights from the last 12 months can be imported. Each imported rating is stored marked as an import and counts towards scores in the same way as a rating made on the form.

To report a rating you believe is fake, or misuse of the rating form, email [email protected].

When a score is shown

A carrier, airport or route shows a score once it has 5 ratings. Below that, the page shows the count and the share of smooth, firm and rough ratings, and the score reads n/a.

Published scores are pulled toward the global mean so that a short run of ratings cannot outrank a large sample. The formula is:

score = (n × mean + 5 × global mean) ÷ (n + 5)

Here n is the number of ratings, mean is the TQI average for the page, and global mean is the average across all ratings. A carrier with 5 ratings averaging 90, against a global mean of 70, scores 80. With 300 ratings at the same average it scores 89.7.

A score built from fewer than 30 ratings is marked provisional in tables, because a small sample can move a lot with one more rating. The threshold was set at 5 while the site is new and may be raised as data grows.

Tables sort ranked rows by score, then unranked rows by rating count. Airports are scored on arrivals only.

Comparing by aircraft, distance and arrival time

Each rating records the carrier, the route and, where known, the aircraft. Aircraft names are grouped into families (all Boeing 737 MAX variants together, all A320-family variants together, and so on), so a family collects enough ratings to compare. A name that matches no family is left out of aircraft comparisons.

Distance is the great-circle distance between the two airports. Short haul is under 1,500 km, medium haul is 1,500 to 4,000 km, and long haul is over 4,000 km. These bands are this site’s own convention.

Arrival time is the scheduled local hour at the destination, grouped as night (00:00 to 05:59), morning (06:00 to 11:59), afternoon (12:00 to 17:59) and evening (18:00 to 23:59). It is only known when a flight lookup matched, so many ratings have no arrival time and are left out of that comparison.

These comparisons show what passengers reported, not the effect of any one factor. Long-haul flights, for example, mostly use widebody aircraft, so a higher score for a widebody type may reflect distance or carrier as much as the aircraft.

Reference data

Runway lengths, surfaces and field elevations come from OurAirports (public domain), extracted in September 2026. Carrier alliance, type, hub and main aircraft families are reference data reviewed in September 2026. Fleets, alliances and hubs change, so treat them as a guide and check the carrier for current details.

Corrections

Ratings are never removed at the request of an airline or airport. A rating is rejected only when its details are invalid, for example a date in the future or an airline code that is not in the list. How any affiliate links would be marked is covered in the editorial policy.