Explainer

How award availability actually works

Award seats work differently from cash fares. They are a separate, rationed inventory that opens and closes without notice, and every program reports it a little differently. This page explains how that works.

An award seat is a separate product

It is not a discounted cash fare. The airline sets aside a limited number of seats for points, on its own terms.

A cash fare and an award seat on the same flight come out of different inventories. The airline decides, flight by flight and cabin by cabin, how many seats — if any — its loyalty program may sell for points, and it adjusts that decision continuously as the flight fills. Two seats at breakfast can be zero by lunch without a single ticket being sold for cash.

That is why award availability has no authoritative public feed. Revenue fares are published because the airline wants everyone to see them. Award space is a loyalty benefit, and most programs we have surveyed treat it accordingly: the search itself sits inside the program’s own tool, often behind a member sign-in or automated-access protection. Anything anyone tells you about award availability — including us — is an observation of one program’s tool at one moment, not a subscription to a source of truth.

This site is built on that premise. Every figure is an observation with a named source and an age, described on where the data comes from. Nothing here is described as live, because nothing anyone shows you is.

Availability comes and goes

Award space is opened and pulled back with no set pattern. A day with no seats today may have them tomorrow, and the other way around.

Airlines open award space when they expect seats to otherwise fly empty, and pull it back when demand firms up. The result is feast-and-famine by date: a Tuesday with four business seats between two weeks that show nothing, or space appearing close to departure when the airline concludes the cabin will not sell out. None of this is a malfunction. It is the inventory behaving as designed — just not designed for your convenience.

This is why searching one date at a time undersells the truth, and why the calendar view shows a whole window with its gaps intact. A day with no seats and a day we could not check are rendered differently on purpose, because they license different decisions: the first says look at other dates, the second says look again.

How wide a window one query can cover is a plan limit, listed on pricing. The honesty of the answer never varies by plan; only the amount of searching does.

What a seat count really means

Programs report availability in at least four genuinely different shapes, and flattening them into one number destroys exactly the information you need.

Some programs state an exact number of remaining award seats. Some stop counting at a threshold, so their “4” means at least four. Some will only confirm that a party of your size can book, with no number at all. And some say nothing about quantity. Each of those is a different claim, and each licenses a different decision — an exact 2 for a party of two is urgent in a way that “at least 4” is not.

We therefore pass each count through at exactly the confidence it arrived with, using one shared vocabulary. The five cases are set out side by side, with the real badges, on how it works.

Itineraries complicate counts further. Availability is often priced for the journey, not the seat map: a connection can be bookable as one itinerary while its individual legs are not (airlines call this married-segment inventory), and a “business” itinerary can put you in economy for a leg. We label mixed cabins as mixed and record itinerary-level answers as itinerary-level, because a count that ignores this is wrong in the way that costs you a transfer.

Every program answers only for itself

“Is there award space on this flight?” has as many answers as there are programs that can book it.

The same physical seat can be released to partners unevenly: one program’s tool shows it, another’s never does, a third wants a different number of points for it. Which partners see what, and at what price, is a commercial arrangement between airlines that no search tool can see behind.

So a search of one program tells you what that program would charge and what that program can see — nothing more. That is why every result here is attributed to the single program that reported it, and why we never merge programs into a blended answer: the disagreement between them is not noise to smooth out, it is the comparison you came for.

Observations age, and the airline's page settles everything

Between the moment a program answered and the moment you book, the answer can change.

Award inventory can disappear between an observation and a booking attempt — that sentence appears on every program page on this site because it is the single most expensive thing to forget. Every result therefore shows its age, on two separate clocks: when we fetched it, and when the source itself says it last checked. Where only our fetch time exists, we say so rather than dressing our clock up as the airline’s.

The booking always happens on the loyalty program’s own site, and that page is the authority. If it disagrees with us, it is right and we are stale — and if you are moving points to book, confirm the seat there first, because a transfer cannot be undone.

Three kinds of nothing

An empty results page is ambiguous unless the product says which nothing it means. We distinguish all three, everywhere.

Confirmed empty

The program answered in full and reported no award seats for that route, date and cabin. This is a real result. You can act on it, and an alert can watch for it changing.

Unavailable

The program did not answer, or answered incompletely. Nothing can honestly be said about its seats, so nothing is. The search is marked partial rather than quietly rendered as empty.

Not checked

The program is not enabled in this product, so we did not look. The absence of its options from your results is a statement about us, not about its seats.

Today, “not checked” applies to every program: none is enabled for production searching, and the coverage page states the exact position of each one — including which are blocked by permission, which by the airline itself, and which by both.

The mechanics of a search itself — cost preview, caching, parsing, the two clocks — are a separate page.

How a search runs, step by step