How to Read Flight Performance Data
A short, practical guide to the BTS On-Time Performance dataset that powers every page on PlainFlights.
Source: Bureau of Transportation Statistics Reporting Carrier On-Time Performance
What "on-time" actually means
Throughout PlainFlights, "on-time" follows the federal definition used by the Department of Transportation in its monthly Air Travel Consumer Reports: a flight is on-time if it departs or arrives within 14 minutes of its scheduled time. Anything 15 minutes or more late is counted as a delay. There is no concept of "slightly late" — it is binary.
This matters because two airports with similar on-time percentages can feel very different in practice. An airport that is technically on-time but consistently 10–14 minutes late will hand its passengers many more missed connections than an airport that is on-time within 5 minutes. The BTS on-time percentage alone does not capture that nuance, which is why every airport page on PlainFlights also shows the average delay in minutes — the second number you should read.
The five BTS delay categories
When a flight is delayed by 15 or more minutes, the airline assigns the delay minutes to one or more of five mutually exclusive cause categories. The airline makes this assignment using DOT guidelines, so the categories reflect the carrier's view of why the flight was late, not an independent audit:
- Carrier delay — within the airline's control: maintenance, crew scheduling, fueling, aircraft cleaning, baggage loading, cabin preparation.
- Weather delay — significant meteorological conditions that prevent safe operations: thunderstorms, snow, ice, fog, wind shear at origin or destination.
- NAS delay — National Aviation System factors: ATC flow restrictions, airport operations limits, traffic volume, runway closures.
- Security delay — security-related events: terminal evacuations, security breaches, passenger re-screening.
- Late aircraft delay — cascading delay caused when the previous flight using the same aircraft arrived late, propagating to the next segment.
On any airport, airline, or route page, the delay-cause breakdown answers a different question than the on-time percentage. The percentage tells you how often flights run late. The breakdown tells you why. For travel planning, the cause matters: weather and NAS delays are largely outside the airline's control and are usually shorter; carrier and late-aircraft delays tend to cascade through the day at busy hubs.
How to read an airport page
Each airport page opens with three numbers: total flights tracked, the on-time percentage, and the average delay in minutes. Read them together. A 78% on-time rate sounds reasonable on its own, but combined with a 47-minute average delay it tells a very different story than the same percentage paired with a 22-minute average.
Below the headline metrics, each airport page also shows the cancellation rate, the leading delay cause, and the busiest months. The seasonal table is one of the more useful sections because it reveals whether the airport has predictable bad windows (summer thunderstorm months in the Southeast, winter storm months in the Northeast) that you might want to plan around.
How to read an airline page
Each airline page shows the same three headline numbers as an airport page — total flights, on-time rate, average delay — applied across that carrier's entire BTS record. The most informative section is usually the delay-cause breakdown, which reveals whether a particular carrier struggles primarily with in-house issues (carrier and late-aircraft delays) or with external factors (weather and NAS).
When two airlines look similar on the headline on-time percentage, the cause split is often what distinguishes them. A high carrier-delay share suggests operational fragility that compounds with bad weather. A high late-aircraft share usually means the airline runs a tight schedule with little recovery slack.
How to read a route page
Each route page aggregates every reported flight on a specific origin–destination corridor. Beyond the standard on-time and delay numbers, the most actionable signal is the count of carriers — routes with more competing carriers tend to have more total flights per day, which means more rebooking flexibility when something goes wrong. A route served by a single carrier with three flights a day offers very little fallback.
The five BTS delay categories side-by-side
Bureau of Transportation Statistics requires reporting airlines to attribute each delayed flight to one of five mutually exclusive categories. Understanding the distinction is critical when comparing two airlines or two airports — what looks like a poorly run airline may actually be operating in a high-weather or high-ATC-congestion environment outside its control. Conversely, a high "carrier delay" share is a strong signal of operational quality issues that the carrier itself owns:
| Category | Definition | Typical share of total delays |
|---|---|---|
| Late aircraft | Cascading delay from the same aircraft's earlier flight | ~30-40% |
| Air carrier | Maintenance, crew issues, fueling, baggage | ~25-30% |
| National aviation system | ATC volume, ground holds, airspace constraints | ~20-25% |
| Weather | Conditions severe enough to ground or divert flight | ~5-10% |
| Security | Evacuation, breach, screening backup | <1% |
Shares approximated from BTS Air Travel Consumer Report delay-cause summaries (bts.gov/topics/airlines-and-airports). Actual values vary by year, carrier, and airport.
Worked example: comparing two airlines on the same route
Suppose you are comparing Airline A and Airline B on the New York-LGA to Chicago-ORD route across the most recent reporting year. Airline A shows a 78% on-time arrival rate; Airline B shows 71%. The headline suggests Airline A is clearly more reliable. But pulling the delay-cause breakdown changes the picture: Airline A's late-aircraft share is 35%, carrier delay 28%, NAS 27%, weather 9%. Airline B's late-aircraft share is 22%, carrier delay 50%, NAS 21%, weather 6%. Airline B's much higher carrier-delay share reveals a structural operational issue that Airline A does not share. Even though Airline A's on-time rate is higher, Airline B's actionable problem is concentrated in a single category — which means a single operational improvement could close most of the gap. This is the kind of insight that the headline on-time rate alone hides.
What the data is not
BTS on-time performance does not tell you anything about onboard experience, seat comfort, baggage handling quality, in-flight service, or pricing. It is a punctuality and reliability dataset, not a customer-satisfaction index. A carrier with the best on-time numbers may still be the carrier you do not want to fly for entirely separate reasons. PlainFlights does not collect or publish those other dimensions — they live in different federal datasets (DOT consumer complaints, baggage and tarmac-delay reports) and in private review platforms.
