How to Pick a Connecting Hub Using Real Wait-Time Data

Most itinerary tools let you choose your connecting hub. Almost nobody uses that lever deliberately, which is a waste, because the difference between the options is measurable and it’s sitting right there in the data.

Here’s the case study I use: you’re flying transatlantic into the US with an onward domestic leg. Atlanta, Charlotte, Dallas Fort Worth, Detroit, and Houston all offer it. Which one?

The customs numbers decide it

For an international arrival, the checkpoint is irrelevant — you’re not going through TSA on the way in. What you’re going through is the federal inspection facility, and the spread is enormous:

Hub	Immigration peak	Worst hour
ATL	75 min	Sunday 3 PM
DFW	53 min	Sunday 4 PM
DTW	53 min	Friday 4 PM / Sunday 4 PM
CLT	52 min	Sunday 4 PM
IAH	50 min	Friday 3 PM / Sunday 4 PM

Atlanta is 22 minutes worse than the next one down. That’s not a rounding error — that’s a connection you make versus one you don’t.

Houston is the surprise winner here. IAH has a reputation as a difficult airport, largely earned on the departure side where its TSA peak hits 60 minutes on Sunday at 6 AM. But its customs hall tops out at 50 minutes, the best of the five, and its midweek numbers are better still: Tuesday caps at 30.

Flat beats low

Now the domestic case, where the checkpoint does matter — you left the secure area, or you’re starting your trip at the hub.

Charlotte’s peak is 25 minutes. Dallas Fort Worth’s is 32. On peak alone, Charlotte wins. But look at the shape:

DFW, worst hour per day: 32, 32, 29, 32, 29, 29, 28. Seven days, an eight-minute spread.

CLT, worst hour per day: 22, 15, 19, 25, 24, 24, 18. Seven days, a ten-minute spread, with the worst days landing Thursday and Friday afternoon.

DFW is higher but flatter. And when you’re connecting, flat is what you want, because you don’t get to pick your hour — the schedule picks it for you. An airport that’s reliably 30 minutes is easier to plan around than one that’s usually 18 and occasionally 25 at exactly the hour your inbound lands.

Budgeting a consistent 45 minutes beats gambling on a number that moves.

The quiet-slot check

One more filter. Charlotte logs 35 quiet hourly slots at 5 minutes or under. Dallas Fort Worth logs 23. Atlanta logs 34.

None of those are good numbers. Compare them to Midway’s 127 or Dulles’s 119 and you can see the category difference: mega-hubs don’t have empty hours. They have a floor, and the floor is somewhere in the low teens.

So don’t route through a mega-hub expecting to find a quiet moment. Route through one because the connection works, and budget for the floor.

The Atlanta defense

That is harder on Atlanta than it deserves, so to be fair: ATL moves more people than any airport on earth and it does it with a plenum-and-train layout that gets you between concourses faster than almost anywhere. Its TSA peak of 45 minutes at 9 AM on Monday is bad, but the concourse-to-concourse transfer time is short enough that a 60-minute connection there is more realistic than a 60-minute connection at, say, O’Hare.

The problem is specifically the international arrival. Seventy-five minutes in the hall at 3 PM on a Sunday, and then you still have bag recheck and a train ride. That’s the scenario where I actively re-route.

The rule I use

Three questions, in order:

1. Am I clearing customs here? If yes, immigration data outranks everything. Pick the lowest peak at the hour you’d actually land.
2. Is my inbound landing in the hub’s worst window? Check the specific hour, not the daily average. Most hubs concentrate their pain in three or four hours.
3. How flat is the profile? If two options are close, take the flatter one. Predictability compounds across a multi-leg trip.

That’s it. It takes about ninety seconds on two airport pages and it has saved me more missed connections than any amount of buffer padding.

What about the coastal gateways?

Avoid them for connections when you have a choice, and the numbers make the case cleanly.

San Francisco peaks at 79 minutes for customs on Sunday at 4 PM. O’Hare at 78. JFK at 77 on Sunday at 3 and 4 PM. LAX at 77 on Sunday at 3 PM. Newark at 76. Atlanta at 75. Miami at 74 on Sunday at 3 PM.

Against Charlotte’s 52, Dallas’s 53 or Houston’s 50, that’s a 22-to-29-minute penalty at the immigration hall alone — before you’ve claimed a bag or re-cleared security.

And the security side compounds it at some of them. JFK logs zero hourly slots at five minutes or under. LAX logs 10. Those checkpoints never empty, so the re-clear step costs full price no matter what hour you land.

The one argument for a coastal gateway is schedule: more international frequencies, more onward options if something goes wrong. That’s a real advantage on a trip where a missed connection is expensive. It’s not an advantage on an ordinary one.

You can compare directly on the JFK immigration wait times page.

Does the arrival hour change the answer?

More than the airport does, in some cases.

Every US gateway peaks between 3 and 5 PM, because that’s when the long-haul banks land — Charlotte 4 PM, Dallas 4 PM, Denver 4 PM, Houston 4 and 5 PM, Atlanta 3 PM, JFK 3 and 4 PM, O’Hare 4 PM, Miami 3 PM. Fourteen airports, one window.

Which means a mid-morning arrival at almost any of them clears faster than a mid-afternoon arrival at the best of them. Charlotte on a Tuesday clears in 37 minutes. Atlanta at 10 AM on a Wednesday is well under its 75-minute Sunday peak.

So the ordering is: pick the hour first, the hub second. If both are available to you, an inland hub in the morning is the fastest re-entry into the country you can book.

CBP publishes its own airport wait-time data if you want to cross-check a specific routing.