The Design of Airline Route Networks

文章要点：

• United, American, and Delta airlines have similar offerings, leading to an oligopolistic competition.
• Smaller airlines differentiate through unique models and risk tolerance.
• United Airlines emphasizes network depth to capture niche markets.
• American Airlines focuses on domestic connectivity and cost efficiency.
• Delta Airlines prioritizes brand loyalty and operational efficiency over expansive routes.
• The personal data privacy issue and solution by Incogni is highlighted.

---

United, American and Delta airlines offer almost exactly the same thing. Not exactly, but almost. And they do this on purpose. When one of them eliminates change fees, they all eliminate change fees. When one of them announces free Wi-Fi, they all announce free Wi-Fi. When one of them ups bag fees, they all up bag fees.

They each offer pretty much the same legroom, pretty much the same snacks, pretty much the same lounges, pretty much the same reliability. Smaller airlines don't do this. They're strategically motivated to differentiate through offering a low-cost model like with Spirit or a unique product like JetBlue Mint. But the bigger US airlines compete in one of the most textbook examples of an oligopolistic industry.

And in oligopolistic industries, there is strategic incentive to converge on product offering. If one player makes a change, they all follow suit so that none of the three gains a meaningful differentiation. It's also mutually harmful for the three airlines to compete via price, so they largely don't and allegedly use algorithms to standardize fares between each other.

So with disincentive to compete on both product and price, that leaves one major arena where the three airlines can compete: Network. Subtly, each of the three largest US Airlines has carved out a small niche for themselves through the nuances of the design of their route networks.

From a high level, they look more or less equivalent. But by digging in, one can see the subtle strategic differences between the three, the ways they attempt to claw out a competitive edge against each other.

For example, there's really only one airline among the three that would fly this route: Anchorage, Alaska to Washington D.C. Of course, that's in part because it ends up in D.C., a United hub. But it's not about the literal start or endpoint of this route, but rather the characteristics of this connection.

It's sort of weak. Prior to launching this route, United already operated flights to the last Frontier from each of its mainland hubs, minus Los Angeles. That meant that they already flew from Chicago to Anchorage. Chicago is a perfect connecting hub for passengers headed from the east coast to Alaska.

It's right on the way, and almost every midsize or larger airport on the east coast is connected to Chicago via United. D.C., meanwhile, is not a good connecting hub to Alaska. It's out of the way from almost every airport, and it sits almost 3,400 miles or 5,500 kilometers away.

That puts it towards the edge of the range of the narrowbody 737 Max 8 assigned to the route. The closer a flight gets to the limits of an aircraft's range, the tougher the economics tend to be due to potential weight restrictions, limiting seat sales or cargo potential. As in overall fuel efficiency gets worse towards the extremes of range.

So operating an Anchorage flight from D.C. offered worse economics and a worse geographic connecting point. And yet, in 2024, United Airlines launched the flight anyway, a 7 to 8 hour daily nonstop operating throughout the summer season.

Now what United was likely banking on, at least in part, was the local market. The hope was that there were a good number of D.C. area residents, and to a lesser extent Alaskans, who were willing to pay a fair premium for the convenience of a nonstop. If there were enough, it would offset the worse economics of the flight.

And there must have been as this flight is now back for the 2025 summer season as well. A summer seasonal narrowbody domestic flight might appear unremarkable, but it's really a uniquely United route based on the level of risk tolerance it required to launch.

American and Delta tend to be conservative in their route planning, whereas United believes in their ability to induce demand through offering more convenient connections. United really competes on network depth. Of the big three US Airlines, it flies the fewest daily flights and carries the fewest annual passengers yet serves the highest number of destinations.

They spread their operations out and focus on breadth versus depth. This sometimes manifests quite objectively when United flies to places American and Delta do not: Puebla, Mexico; Farmington, New Mexico; Laramie, Wyoming; Devil's Lake, North Dakota.

It also manifests through year-round service to destinations their competitors only fly seasonally: Edinburgh, Scotland; Gunnison, Colorado; Auckland, New Zealand. Each of these three sees seasonal tourism surges yet still has reasonable year-round demand. United captures that year-round demand. Delta and American compete for the seasonal surge.

United's network depth takes its most extreme form outside North America. American Airlines serves 26 different long-haul international destinations, 18 of which are in Europe, leaving one flight to Qatar, one to India, three to East Asia, two to Australia, and one to New Zealand. Delta flies to 41 long-haul destinations, the difference largely made up of additional summer-only flying to key markets in Europe, plus a six-destination Africa market.

United, meanwhile, flies to 58 different long-haul destinations. Much of this is made up of depth in already strong markets. Delta and American fly to Tokyo, but United is the only one to fly to Osaka, a top destination for American tourists in Japan and also a strong origin market for US-bound passengers in New Zealand.

Delta and American both fly to Auckland seasonally. Yet United is the only to fly nonstop from North America to Christchurch, the largest city on the country’s less populous but more picturesque South Island. With a route like this, the airline is banking on its ability to induce demand.

Patrick Quayle, the United executive overseeing network planning, has discussed in interviews the difficulty in proving the viability of a potential route like Christchurch through quantitative data.

When starting a route domestically, much of the decision can be made by looking at traffic pattern data that shows, say, that enough people are flying from D.C. to Anchorage via connections that it’s worth launching a nonstop. But it’s unlikely that the numbers directly demonstrated that for a relatively minor long-haul market like Christchurch, because tourists to New Zealand were historically likely to book a ticket just to Auckland.

From there, some might rent a car, some might stay for a few days, and book a separate ticket to Christchurch. Some might even book a ticket to a different south island airport like Queenstown. So the theory behind a route to Christchurch went well beyond just capturing the customers that previously booked a connecting itinerary, but rather to induce people to form their plans around the convenience of the flight and fly straight to Christchurch.

This is certainly primarily a leisure flight; there’s not much business demand to Christchurch, and leisure travelers tend to be fairly flexible with their plans. United has had decent success in the past at inducing demand on leisure-focused long hauls. Other examples would be their flights to Tahiti or Cape Town.

So in some ways, United’s competitive advantage centers around their risk tolerance. Christchurch didn’t have the data to prove it’d be profitable. The airline subjectively considered it worth a shot. If they’re right, United can essentially monopolize the market before American, Delta or another competitor can enter.

And recently, United has pushed to new extremes in developing far-flung markets. It started as a solution to a problem. You see, United has long operated a rather unique small hub in the American territory of Guam, a mere three-hour flight from Tokyo, meaning they have crew and narrowbody 737 aircraft stationed on the far side of the Pacific.

Previously, the financial motive behind such a far-flung atypical hub was focused on capturing the large inbound tourist market from Japan to Guam. Since the COVID pandemic, though, the territory's tourism numbers have fallen dramatically, in part due to the weak exchange rate between the yen and the dollar, making vacations in the US territory much more expensive for Japanese people.

Rather than redistributing aircraft towards the mainland and shrinking the hub, United opted to use this excess capacity to develop a deeper route network into Asia. Specifically, they’re turning Tokyo Narita Airport into a mini hub with new narrowbody routes to Cebu, Philippines; Kaohsiung, Taiwan; and Ulaanbaatar, Mongolia.

These flights are each operated using Guam-based aircraft, yet are timed to leave Tokyo in the early afternoon, providing quick connections from the airline's inbound long-haul flights from the US, which all arrive within a few hours of each other. Then the 737s return from their destinations the following morning to time with the next day’s US-bound long-haul bank.

With each of these routes, no United partner airline operates a flight at a time that allows for same-day connections via Tokyo, meaning passengers would often have to overnight in Japan or more realistically, fly with another airline. Now, United is able to advertise the fastest itinerary between the US and these destinations, and they can likely even win some additional customers through offering an itinerary fully on their own metal rather than on a foreign airline.

US passengers might be unfamiliar and apprehensive of this, but even though some of these routes have not yet flown, the strategy must be proven promising since the airline is already doubling down on it with newly announced routes from Hong Kong to both Bangkok and Ho Chi Minh City.

Like the Tokyo routes, these will be timed for an easy connection to and from the airline's Hong Kong long hauls. Yet in this case, they’re not being operated by a Guam-based aircraft, rather a valuable 787 that could otherwise be operating lucrative long hauls.

Really, the more you dig into it, the more United’s international network design is explained through the competitive weaknesses of their partner network over in Europe, the most important long-haul market for US Airlines.

United has a major transatlantic joint venture with the Lufthansa Group, meaning the airlines have essentially formed a jointly owned business and share revenue between all their transatlantic operations. In practice, this means they work together to grow economies of scale by forming networks that integrate together well so that the joint venture is more competitive overall.

But American’s biggest transatlantic partner is British Airways in London, whereas Delta connects passengers primarily through Paris and Amsterdam via Air France and KLM. The biggest Lufthansa Group hubs, though, are Frankfurt, Munich, and Zurich, far further into Europe.

When flying from the US, that makes them less competitive, both costlier for the airline and more time-consuming for the consumer. But it's that very strategic shortfall that explains the vast majority of United's nonstop flights to Europe sit to the southwest of the Lufthansa Group's primary hubs, as this is the area of Europe least effectively served by them.

A United traveler headed from Denver to Faro, Portugal, for example, would have to sit through an almost 16-hour itinerary when connecting through Frankfurt onto Lufthansa, whereas with the nonstop from Newark, they can fly there in about 13 hours, the fastest itinerary offered by any airline.

While it’s certainly true that American travelers are more likely to visit Western Europe, explaining some of the disproportionality, United has tried routes to the north and east, where the Lufthansa hubs would be more comparatively competitive. They used to fly to Hamburg, but that stopped in 2018. They used to fly to Prague, but that stopped in 2019 after a single disastrous summer.

They used to fly to Bergen, but that too lasted only for the summer of 2022. But despite these few stumbles, United’s international strategy appears remarkably effective. In Tenerife, Faro, Malaga, Bilbao, Palma, Dubrovnik and elsewhere, they’ve proven the only airline, American or otherwise, able to make the economics of a transatlantic flight work.

Meaning they have these markets cornered. And increasingly, this globality is expanding to the Pacific, as they’re the only airline offering Trans-Pacific service to Christchurch and Adelaide, and the only US airline serving Bangkok, Ho Chi Minh City, Cebu, Manila, Singapore, Kaohsiung, Osaka, and Ulaanbaatar.

This has set United up for the current moment so well, as the strongest segment of US-origin travel in recent years has been long-haul international leisure travel. And certainly, based on the recent growth of this sector of their network, United has found flying to far-flung destinations worthwhile, expanding their ability to capture more customers and charge more.

It's all so very different from the strategy of American Airlines. They're the US second largest airline by destination count, but by far the largest in terms of passenger count. This reflects the reality of their network relative to United. They skew more towards domestic focus and are less likely to serve niche routes internationally. They just do what they need to.

In fact, they've more or less said this outright to their investors: "In a 2024 our short-haul network is the foundation of value for customers and investors." So while they don’t necessarily find much reason to strongly compete internationally, American does need to offer some global connectivity to meaningfully compete at all with United and Delta.

It's particularly important for winning corporate contracts. American, Delta, and United each compete to sign deals with companies requiring lots of business travel. Think McKinsey, Deloitte, Lockheed Martin, Boeing—really any large business with a dispersed geographic footprint. These contracts are predictable business and tend to include plenty of lucrative premium travel, making them hugely valuable for airlines.

Yet in addition to price, perks, and terms, one of the major things a potential client might look at is global connectivity. Deloitte will be looking at how easily their executive from Dallas might be able to get to, say, Stockholm. American's answer to this more than anything is London.

London is a fantastic European connecting hub because it’s geographically positioned at the entry point to the continent coming from America. And it’s home to British Airways, which offers a huge network of well-timed connecting flights into the continent. British Airways is American's joint venture partner, exclusively, making London a perfect way for American Airlines to serve Europe without having to take on all the risk and capital cost of developing a big network into the continent.

It might not often be competitive to United's nonstops, but it gets the job done. For this reason, American flies a colossal quantity of flights to Heathrow: twice daily from LA, once from Phoenix, five times from Dallas, three from Chicago, two from Miami, three from Charlotte, two from Philadelphia, and four times from New York.

Those are all of American's international hubs, but they even fly daily Heathrow flights from two non-hub airports, Raleigh and Boston. That’s 24 daily flights overall, providing almost 7,000 daily seats with 69 daily transatlantic frequencies overall in the summer. That means more than a third of American’s European capacity is dedicated exclusively to Heathrow.

And the proportions are even more stark during the leaner winter schedule. And this makes perfect sense. In fact, American hardly has another option. They retired a huge portion of their long-haul fleet during COVID, now widely considered a strategic blunder—and production backups at Boeing and Airbus make for long timelines to acquire new aircraft.

So, with the limited widebodies they have, they focus their fleet onto the place with the most onward connectivity. Then beyond London, they focus on the biggest hits. Not a single one of their European destinations does not also receive flights from either United or Delta. They’re not trying to compete on convenience; they’re trying to minimize their costs, their risk, and focus on what they’re good at—short-haul.

An American is very, very good at competing on short-haul. In part just thanks to luck, business came to them. This slide from their investor presentation explains it. The blue dots represent airports where American is dominant, at least based on their own analysis, either through offering the most destinations or flight frequencies compared to their competitors.

This shows that the region of the US where American really most consistently excels is the Sun Belt. Arizona, Texas and Florida are some of the fastest growing states in America, and in each, along with much of the rest of the Sun Belt, American is the dominant airline.

American itself admits to investors that its strategy is centered around cornering the market in certain cities, particularly in the Sun Belt. In its presentation, it brings up the example of El Paso. Delta barely competes there, flying just to Atlanta. United does a little more flying to Denver, Chicago, and Houston. But each of these hubs is fairly far and not in the orientation of the biggest passenger flows.

There aren’t many big destinations to the Southeast, Northeast, or North. American, meanwhile, flies to its hubs in Los Angeles, Chicago, and crucially Phoenix and Dallas. Each of this pair is nearby and on the orientation of the largest east-west passenger flows.

The difference between flying from El Paso to San Francisco via Phoenix vs Denver might seem minor at first, but it represents the difference between a six-hour and four-and-a-half-hour itinerary, which can absolutely be the deciding factor for passengers. And it also means that American’s operating costs for the itinerary will be lower, given fewer flown miles since airlines often converge on price using algorithmic collusion.

This’ll mean on average, America’s profit margins on this segment will also be higher. The compounding effect of these well-located Southern hubs really adds up. So for cities like El Paso, American asserts competitive dominance by offering the largest amount of capacity in the market, providing four daily flights to Phoenix, three to LA, nine to Dallas, and two to Chicago.

This is in stark comparison to, say, Boise. American just doesn’t have the hubs to meaningfully compete in the Northern capital, so it doesn’t bother. It flies there from Phoenix, Dallas, and Chicago with a total of eight daily flights in the summer, providing connectivity for its loyal customers, but not sinking resources into trying to be the dominant carrier in a place where it would struggle to win against United and Delta’s footprint given their Denver and Salt Lake City hubs.

Now, all airlines do this to some extent, devoting a ton of capacity towards markets where they believe they can win thanks to geography. And limiting capacity towards less geographically advantageous markets. American just makes it a more overt and aggressive component of their strategy.

They have found that through sheer scale they can dominate midsize markets like Albuquerque or Oklahoma City or Memphis, making American Airlines the clear choice, especially for the outbound markets since they offer more departures on more efficient routings from much of the Sun Belt to much of the rest of the world. American also extends this strategy south into Latin America, given the advantages of their Los Angeles, Phoenix, Dallas, and Miami hubs.

They’re each geographically suited for the north-south cross-border flows and located in regions of the west with high Latin American populations. American operates to 29 different airports in Mexico, for example; United meaningfully competes with 22 destinations in the country.

It does have a well-located hub in Houston after all. But Delta really only focuses on the major tourist destinations, with nine Mexican airports served. Pretty much the same can be said of the Caribbean. With 28 different relatively small independent states and territories, there is no major dominant region-specific airline. There are minor ones like Caribbean Airlines, but American Airlines manages to rank as the Caribbean’s largest airline overall despite not even being based in the region.

Once again, just dominating a regional market through sheer scale, thanks to its perfect, perfect Miami hub. Of course, then there’s Delta. Strategically, Delta's network sits somewhere between Americans and United’s well. Of course, one airline has to sit in the middle of this spectrum. There’s a good reason it’s Delta.

Delta does not consider its network as core to its strategic advantage. In fact, in their 2024 Investor Day presentation, they didn’t even mention their network. They relegated the two slides about it to the appendix. Much of their presentation focused on what the company believes is their competitive advantage: their brand.

Delta considers itself a premium airline offering a differentiated elevated experience compared to United and American. In prior years, the airline experimented with more concrete notable differences, such as an elevated long-haul economy experience where a passenger received a welcome cocktail and hot towel after takeoff, the selection of both appetizers and main courses, chocolates handed out before landing, and more.

But these went away and never came back post-Covid. This indicates that Delta must not have seen enough of a revenue impact to justify the additional operating cost and the industry’s oligopolistic tendencies. Once again, incentivized product convergence.

These days, Delta Economy Class lacks major objective differentiation in terms of onboard experience compared to United and American; it’s a bit ahead of the curve on rolling out free Wi-Fi (much of its fleet has been refurbished in recent years), but its primary objective, product differentiation, is a consistently high on-time rating higher than United’s and American’s year in, year out.

Customers also often anecdotally report better service. But how objectively differentiated Delta is doesn’t matter much, at least in the short term, because customers clearly believe it is. In recent years, it has tended to beat United and American in profitability, as it points out to its investors.

Delta is able to command a revenue premium thanks to its brand, its product, its loyalty program, its credit cards, but not as much its network. Delta’s network is cost-efficient, serving its own purposes, but it’s not necessarily the most competitive in capturing customers. They less often offer the fastest itinerary and they are firmly America’s third-largest airline by destination count.

Their network is not as vast as United's, nor as geographically efficient as American's. For example, Salt Lake City is a good hub for sure, but United's in Denver is better. Denver has a larger local market, boosting the economics but, and the airport is also located far outside the city with nearly limitless room for expansion, whereas Salt Lake City airport has limited opportunity for growth, both given the configuration of the mountains around and the fact that the area surrounding the airport is already built up.

And perhaps most crucially, Salt Lake City is just not as good of a geographic position. Denver is a perfect location for serving the large east-west passenger flows to smaller Mountain West destinations. And therefore United has a dominant market position into high-yield destinations like Telluride, Aspen, and Jackson, along with market dominance in small destinations like Durango, Grand Junction, or Casper.

Despite their nearby hub, Delta barely even competes in this region, offering connectivity to major destinations but hardly attempting to maintain a dominant position in much of the Mountain West. It’s a similar story with many of the rest of their hubs. Seattle is weaker than San Francisco; it’s a smaller local market, it sees heavy competition from Alaska Airlines.

Perhaps the only benefit is it’s slightly better geographically suited as a Trans-Pacific hub. Minneapolis and Detroit both serve roles better served by Chicago, and Delta just competes directly with American and United in New York and LA. And direct competition almost always leads to worse economics.

Atlanta though, that’s an exception. It is truly the crown jewel of Delta’s network. It’s a dream location for an airline hub. It’s perfectly located to capture massive north-south passenger flows. A ton of people live in the Northeast, a ton of people visit Florida, the Caribbean and Mexico.

And yet Atlanta’s far enough west to be competitive for east-west traffic, offering itineraries competitive with America’s through Dallas to much of the western half of the US and then it sits in the center of the American south, really the sole large region that the airline is able to just simply dominate in terms of connectivity. By and large, Delta’s strength is in their cost management and brand loyalty.

They can be risk-averse and yet still effectively capture customers. Meanwhile, American and United do more of the endless work of continuously tweaking their networks to best serve every single configuration of every single route to, from and between each and every US airport.

One rather concerning thing about living in America is that there are rather weak data privacy laws and a whole lot of personal data that is put out into the public record, often by governments themselves. This wasn’t a huge problem before the Internet, but nowadays it means that people can turn a profit off of making it easier for people to get your personal info. And they do.

It’s called people search sites. Just search your name and address or phone number, and you’ll probably come across a whole lot of sites that have at least some of your personal info. But you do have the power to change that. You have the right to request your data removed from each and every one of these sites. The problem is… it takes forever to do that.

They make it as hard as possible, and you have to do it for every one of the endless sites that have it. But that’s a problem our sponsor, Incogni, fixed. They do that whole process for you. They constantly search the Internet for any of your personal data and go through the entire process of taking it down, including handling rejections.

Not only is this good for peace of mind, but it can also help cut down on annoying spam calls and emails, since it’s these same services that sell your data in bulk. I’ve been an actual paying customer of Incogni for years, so I know it works. And I find the cost well worth it. And it’ll be even more worth it for you, because Incogni is offering 60% off an annual plan when you go to incogni.com/wendover and use the code Wendover.