How a flight alert app works
Five steps sit between a price moving and an email reaching you. Understanding them is enough to tell a useful tool from a notification machine.
A flight alert app queries fare sources continuously, compares what it finds against your criteria, and only writes to you when something matches. Put that way, it sounds simple. What separates a useful tool from a notification machine is everything happening between those three verbs.
The problem it solves
Airlines adjust fares several times a day based on load, season, competition and time to departure. A ticket can shed 30% on a Tuesday evening and take it back by Thursday.
No human sustains that pace for months. Most people watch for three weeks, lose interest, and buy at the price on the day they gave up. Automation does not make you smarter: it makes you present.
The five steps between the price and your inbox
This is the part nobody describes, and it is the part that determines quality.
1. Collection. The tool queries one or several sources — distribution APIs, aggregators, flight search engines. A single source creates blind spots: each one carries different agreements and different fare bands.
2. Normalisation. Responses do not arrive in the same shape, nor always in the same currency. A fare in US dollars compared without conversion against a Canadian dollar budget produces 30% errors — one of the most common faults in the sector.
3. Filtering. The offer is matched against your criteria: route, budget, layovers, baggage, time windows, airlines. The finer the filter, the fewer offers survive — and that is the point.
4. Qualification. Is the surviving price actually good? Without a comparison to the route's history, a tool can only say "the price changed", never "this price is low". The difference is decisive.
5. Deduplication. The same flight comes back on every cycle. With no memory of what was already sent, you receive the same offer ten times and stop opening. This invisible step does more for perceived value than the other four.
Four criteria for choosing
- Multiple sources. A single data source means partial coverage, by design.
- Criteria granularity. Being able to filter on checked baggage and time windows, not just a route and two dates.
- Per-route history. Without a price reference, no tool can qualify an offer.
- Noise control. The right question is not "how many notifications" but "how many did I act on".
Three questions to ask before subscribing
Where does the data come from? A tool that will not say usually has one source.
How is the displayed saving calculated? If the percentage is measured against yesterday's price rather than the route's usual level, it means nothing.
Who earns what when I book? A commission-funded tool is not neutral about when you book. A subscription is more so — not a virtue, an alignment of interests.
What no app can do
Worth stating plainly.
Predict prices reliably. Pricing systems are closed and recalculate continuously; tools displaying an accuracy rate are selling a feeling of control.
Catch mistake fares. They sometimes last hours. By the time a cycle detects one and an email lands, it is usually over — covered in mistake fares: what to do when you find one.
Decide for you. An alert flags a window. If you cannot book within forty-eight hours, the tool will not be worth its price.
The Kestrelia difference
Kestrelia is a pure-subscription monitoring tool: 31 combinable criteria, multi-provider querying, and qualification of every offer against its own route's history.
In practice, step 4 — the one most tools skip — is the heart of the product. You are not alerted because a price moved, but because it genuinely dropped below its usual level, on a flight that meets your constraints.
To put it into practice: how to set up a flight alert in 4 steps. And to place the market's tools against each other, Going vs Hopper vs Kestrelia.