United Goes in for the Kill in San Francisco


At the APEX Expo in September, United CEO Scott Kirby talked about how he likes to enter fights when he has the high ground. Otherwise, he’s not interested. Well, San Francisco is one of those places where United has the high ground, and it now smells blood in the water as Alaska shifts its resources elsewhere. It is going for the kill by adding a ton of new frequencies in the domestic market to a variety of existing destinations… and one new one.

United’s increase follows Alaska’s decrease. Alaska has had to pull resources from elsewhere in order to fund its growth in both Portland (OR) and San Diego. It’s Los Angeles and San Francisco which suffer the most. In Los Angeles, airlines are always beating each other over the head trying to get a leg up, but there are gate constraints for United there.

But in San Francisco, it’s a different story. Alaska remains the number two airline behind United, but it will now end service to Austin, Boston, Burbank, and Newark with Orlando becoming seasonal. This redirection of resources has triggered United to pay very close attention to what it is doing at SFO. It must be thinking if it plays its cards right, it can make Alaska irrelevant to the key local travelers.

United will increase flying in Boston and Orlando, but this isn’t about chasing Alaska out of individual markets. Instead, United is thinking about San Francisco on the whole as it figures out what it can do to further cement its leadership in the region and create more distance between it and everyone else.

This is the S-curve in action right here. The airlines with the highest share of capacity will get an even higher market share because of their utility. United is already by far the largest carrier there, but it sees a chance to accelerate that process even further. This isn’t a new strategy, but it’s like trying to put the final nail in the coffin.

United and its joint venture partners have seen their seat share rise from the 46-47 percent range a decade ago to the 53-55 percent range this year. Alaska, which had reached as high as around 15 percent in 2018, is now back in the single digits consistently for the first time since 2010. Simple math tells you that United isn’t just taking share from Alaska, and that’s not a surprise. It’s also taken a big chunk of share away from American and its joint venture partners. And ULCCs remain a rounding error at the airport.

With this shift in seat share, United has seized its hold on the local market even more. Just take a look.

SFO Local Domestic Passenger Share Over Time

Data via Cirium, includes HA/VX for Alaska and CO for United

EDIT: The chart above and numbers immediately below have been corrected

United was carrying just shy of 40 percent of local domestic passengers to and from SFO until the pandemic. Since then, it has steadily climbed to the 42-43 percent range. The last three quarters have seen United top 46 percent for the first time. Alaska? Well, like its capacity, its share of local passengers peaked near 25 percent in 2018. It’s now between 16 and 18 percent.

In other words, Alaska has seen fluctuations, but United has seen outsized growth in its local share. There can be noise here — airlines can always tweak revenue management systems if they want to take more local or more connecting traffic — but even if there is some of that in here, United is clearly growing its local share through network strength as well.

Looking forward, the airline will grow even further in the market. There are six markets that will get two additional daily frequencies. Then there are another 14 markets that will get one extra. And lastly we have a new market down in Carlsbad which will start with 2x daily.

New United SFO Frequencies

Naturally United will add more flights to hubs and help build connectivity. But it will also increase intra-West flying and Hawaiʻi, the two most important local markets where Alaska still has a strong presence.

In the end, this will make United more appealing to local travelers, and it will make it harder for Alaska. Thatʻs not to say I fault Alaska for what itʻs doing focusing on other markets where it either is a clear number one or has a chance to fight for that. Itʻs just that Iʻm sure Alaska would rather United didnʻt see such a good opportunity to keep growing at SFO.

Get Cranky in Your Inbox!

The airline industry moves fast. Sign up and get every Cranky post in your inbox for free.

Brett Avatar

23 responses to “United Goes in for the Kill in San Francisco”

  1. Mr Eric Avatar
    Mr Eric

    I’m shocked United’s local share at SFO is so low. I didn’t realize this market was so fragmented.

    1. Jake Avatar
      Jake

      This seems incorrect, unless it’s including OAK/SJC? That would imply >50% of SFO domestic O&D traffic is AA+DL+WN+B6…which can’t be right.

      1. Alex Hill Avatar
        Alex Hill

        It almost certainly is right. See my other post: it’s the simple fact that nearly every passenger on the (non-AS) competition is local, whereas UA has connecting traffic. It’s always the case at a hub. (And you didn’t include AS, the second largest domestic airline at SFO!)

        UA only has 50% of the seats, so they can’t possibly have a higher share of the local traffic than that given that they carry connecting traffic. UA is the dominant carrier at SFO, but it’s not a CLT-like fortress hub.

      2. Brett Avatar

        Jake and others, honestly, I double checked the data and now I can’t figure out what I did before. It has been updated with both Alaska and United having a higher share. I don’t know what happened but directionally it’s simliar.

        1. Mr. Eric Avatar
          Mr. Eric

          Thanks Brett. Those new figures align better to what I was expecting UA to have at SFO.

    2. Alex Hill Avatar
      Alex Hill

      It’s always the case at a hub that the local market share is dramatically less than the hub airline’s seat share. That’s simple math: for the non-hub airlines, essentially every passenger is local. For hub airlines, some fraction of their passengers are connecting.

      If you have a 90% seat share and 80% of your passengers are connecting (which I think is a pretty good approximation of AA at CLT), you “only” have a 64% local traffic share. If you have a 50% seat share and 50% of your traffic is connecting, you only have a 33% local traffic share (assuming no connections on other carriers for simplicity).

      And Cranky’s numbers are domestic traffic only. Since UA has a lot of domestic traffic feeding their international operation at SFO, that’s probably part of it.

      1. emac Avatar
        emac

        CLT vs. SFO is a good example.

        Something that’s interesting here: UA has only two (going up to four) daily flights to DFW — two! AA has NINE.

        SFO is a destination that attracts capacity from competing airlines’ hubs. CLT less so. So UA is giving up local share to AA in DFW (maybe PHX, PHL and so on too), DL in ATL and MSP and DTW. CLT isn’t as big so it doesn’t attract competitor capacity the same way (CLT-IAH is 5x AA 3x UA, CLT-DTW is 7x AA and 4x DL).

    3. Daniel Paulling Avatar
      Daniel Paulling

      How do you define fragmented? Here’s the 2024 fiscal-year data:

      United Airlines, 46.8%
      Alaska Airlines, 10.4%
      Delta, 8.1%
      American Airlines, 6.8%
      Southwest, 4.0%

      1. Brad Avatar
        Brad

        Those percentages add up to 76%, who is hauling the other 24% of local traffic?

      2. LRK Avatar
        LRK

        Is this total seats, or local traffic?

        1. CraigTPA Avatar
          CraigTPA

          Total seats – his figures are from Wikipedia, which come from a SFO fact sheet for 2024.

          Original here: https://www.flysfo.com/about/about-sfo/sfo-fact-sheet

          It includes domestic and international, local and connecting.

  2. Avelo Jello Avatar
    Avelo Jello

    Brett, what was the purpose of the Virgin acquisition if in the end Alaska was going to shrink both LA and SF to fund growth in their old focus cities. Wouldn’t it been cheaper to just gone at it with their own fleet plan?

    1. SEAN Avatar
      SEAN

      I’ve asked a similar question here more than once & the answer I keep getting is… to keep JetBlue from growing on the west coast by denying them Vergin America. That denial in the end was unfortunately successful.

      1. Avelo Jello Avatar
        Avelo Jello

        In hindsight, it feels like B6 would have imploded in a similar fashion as they did in BOS and JFK – just barely relevant in highly competitive cities dominated by legacy carriers. Meanwhile, AS would have thrived throughout the rest of California and dominated other major metros with $1B in the bank.

        1. CraigTPA Avatar
          CraigTPA

          JetBlue is quite a bit more than “barely relevant” at BOS and JFK, and they certainly haven’t “imploded” in either market.

          It’s entirely possible JetBlue could have made a much more effective run at an intra-West network if they had established substantial operations at SFO and LAX and pulled the plug on LGB sooner than they did. It’s also entirely possible that B6 and AS would have gotten into an intra-West war of attrition, which could have even led to a merger between the two.

          (And in turn, that merger might have finally convinced Alaska to do what the Network Three have already done and realize that they can’t be dependent on either Boeing or Airbus for their larger single-aisle airplanes, and Alaska wouldn’t be in the position of having to cut flights out of SFO right now because Boeing can’t deliver airplanes on time.)

          Of course, it’s also entirely possible that the Virgin purchase would become a millstone around JetBlue’s neck and pull the entire airline down.

          Fun to think about. But in the end, the defensive move by AS was probably the right decision at the time.

    2. Chris Avatar
      Chris

      It was to keep VX away from B6 so that B6 could not get a stronger foothold on the west coast.

    3. Brett Avatar

      Avelo Jello – This is ancient history. They certainly didn’t think 10 years ago that they were going to buy Hawaiian and refocus their growth in different cities after a pandemic. At the time, it was a way to keep Jetblue out of the west coast and acquire gate space in LA and SF. That was part of the airline’s strategy to grow its credit card portfolio in California by becoming more relevant. That much worked.

  3. Jason Avatar
    Jason

    I would really like to see the total increase in seats that UA is providing to these cities. For example to Kona they are going From 2 flights 1 737-8 and 1 777-200 HD to 4 737s that is a much smaller capacity change than adding 2 flights would imply. Also, for ORD to SFO over the years I have seen United increase the number of flight many times just to eventually pull them back. If I remember correctly when VX added flights to ORD UA went to 12 flights a day as a response.

    1. Brett Avatar

      Jason – Looking at July 2026 vs July 2025, Kona is at 23,901 seats up from 17,094, so it’s a hefty increase. Every single one of the listed markets in the chart above has an increase in seats with the exception of O’Hare which is actually off just 1.1%. But more frequency means more choices for customers, so that’s still a win for travelers.

      As of now, the North American markets that are showing down in seats by more than 10% are Anchorage, Salt Lake, Aspen, Puerto Vallarta, Kalispell, St Louis, Cabo, Cancun, Monterey, Monterrey, San Antonio. So it’s a lot of leisure, the Monterey sacrificial lamb, and also, screw you St Louis and San Antonio.

  4. Jimmy Avatar
    Jimmy

    United has also changed its pricing strategy on SFO nonstops that don’t have another airlone on the route. UA used to price a route like SFO-MFR at $400-500 round trip. Now it’s under $200, which is low enough to get people to switch from driving 5-6 hours. It’s changed my family’s behavior.

  5. NedsKid Avatar
    NedsKid

    If you’re going to San Francisco, be sure to wear a tulip in your hair….

  6. abcdefg Avatar
    abcdefg

    Couple thoughts on various threads in here:
    -It’s fairly common for UA to be at a schedule disadvantage between its hubs and OA hubs. I chalk this up to UA hubbing ~top 10 markets only, where their hub city is a destination for a vast majority of the country. CLT/MSP/DTW/SLC/DFW typically have more service into UA hubs like SFO than UA has into those cities. Not an absolute rule, but it can certainly account for share discrepancies when only domestic originating pax are considered

    -Alaska’s strategy with VX was to strengthen their position holistically in California. The acquisition certainly did that and it raises the question of how markets now get defined and what service level has to be offered in related cities (markets). My recollection is that to properly compete in all west coast markets AS needed more utility out of SFO and other CA cities which VX brought, even if it meant being #2 in places like SFO because the utility was sufficient around the entirety of the west, vs UA where everything touched SFO/LAX/DEN only. It would appear that strategies and market power have evolved to where the desire is to own the airport and not the region, with HA bringing dominance at HNL, focus on SEA retention, PDX expansion and buildup of SAN.

    -Coupled with that above point it makes one wonder what the point of no return is for marketshare in a 2 player primary market. AS got to 25 vs ~38 for UA but never could get closer. Is it within 8 points of parity? 5? 3? Has ramifications for SAN, ORD, BOS and a few other contested markets in the country.

  7. David C Avatar
    David C

    Once United started going double daily on so many pan-pacific markets, the power of the connecting traffic must be a real bolster for the health of the domestic flights’ economics. They seem to be digging in at SFO and ORD…and that’s a LOT of gates at Terminal B north project in Houston.

    They aren’t playing around.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Cranky Flier