How would you design the upgrade process if you were in charge?
Apr 9, 2016, 3:05 pm
#31
Join Date: Dec 2015
Location: MSP/ORD
Programs: UA 1MM/GS, Marriott Lifetime Platinum
Posts: 318
Put differently - who, exactly, flies regularly on business but actually has the flexibility to book front cabin at issuance domestically?
This isn't just me - this is United's bread-and-butter customer base.
Apr 9, 2016, 7:07 pm
#32
Join Date: Sep 2014
Posts: 25
If it were idiosyncratic to my (extremely large) employer, that would be one thing. But doesn't almost every large employer use a corporate travel system with a similar restriction?
Put differently - who, exactly, flies regularly on business but actually has the flexibility to book front cabin at issuance domestically?
This isn't just me - this is United's bread-and-butter customer base.
Put differently - who, exactly, flies regularly on business but actually has the flexibility to book front cabin at issuance domestically?
This isn't just me - this is United's bread-and-butter customer base.
Apr 9, 2016, 7:18 pm
#33
A FlyerTalk Posting Legend
Join Date: Apr 2001
Location: PSM
Posts: 69,232
And after back-end discounts (corp rates are discounts off of base fare in general) the bulk sell closer to the lower cost anyways.
Also keep in mind that yield optimization doesn't necessarily mean selling all the seats all the time. It means getting the most revenue for the aircraft as often as possible. So, while many more people might pay $4k instead of $7-8k that also isn't necessarily what maximizes revenue. If the company can sell half as many seats at $8k and then have others "left over" for day-of-travel upgrades, award redemption or other use then it might be more beneficial than selling more seats at the lower rate.
Also keep in mind that yield optimization doesn't necessarily mean selling all the seats all the time. It means getting the most revenue for the aircraft as often as possible. So, while many more people might pay $4k instead of $7-8k that also isn't necessarily what maximizes revenue. If the company can sell half as many seats at $8k and then have others "left over" for day-of-travel upgrades, award redemption or other use then it might be more beneficial than selling more seats at the lower rate.
Apr 9, 2016, 7:28 pm
#34
Join Date: Jan 2013
Location: Hawai'i Nei
Programs: Au: UA, Marriott, Hilton; GE
Posts: 7,132
Revenue managers at UA are not stupid. They understand consumer behavior, and the fact is: Variable Ratio schedules of reinforcement lead to high rate behavior. That is exactly the type of schedule we are all on when it comes to receiving upgrades on UA. Higher status flyers may have a slightly different ratio schedule than lower status flyers, but the principle continues to operate to keep us salivating (apologies to Pavlov's dogs).
Last edited by 747FC; Apr 9, 2016 at 7:30 pm Reason: spell check
Apr 9, 2016, 8:10 pm
#35
FlyerTalk Evangelist
Join Date: Jul 1999
Location: Ewa Beach, Hawaii
Posts: 10,907
Apr 10, 2016, 8:51 am
#36
Join Date: Jun 2006
Location: San Francisco
Programs: UA 1K, Citi Prestige, AMEX Platinum, SPG Gold
Posts: 720
Ideally the upgrade system would be strategyproof (in the sense that United would only lose revenue or expose themselves by deviating from the published rules), as transparent as possible, and would still offer some benefit to elites. Right now, the problem is that if United deviates from their own published rules--e.g., bypassing 1Ks on the upgrade list, or prioritizing miles+cash upgrades over GPUs--then they actually make money or break even. Either way they have little incentive to fix the problem, and when there is a perceived problem they generate a lot of ill will.
So for transparency, what I would do is consolidate into one upgrade list available well in advance and make it public. To address the privacy issue (which frankly is not great even under the status quo), instead of people's initials it should be part of the reservation number of something.
Now to make the upgrades strategyproof, you want the same pricing available to everyone. So first of all, United should restore the ability of people to buy up to premium cabins without a change fee. Second of all, the price of TOD upgrades should be publicly available (i.e., you can get it on the flight status page without logging in). Moreover, the effect of TOD on PQMs and PQDs should be disclosed before purchase (and ideally they would all be some sort of buy-up that counts towards PQD).
Many strategyproof techniques involve some sort of auction. Thus, to give elites some advantage you should be able to choose how to pay for an upgrade after getting the upgrade. For example, if a TOD is available for $100, I should be able to buy it, then, if desired, substitute an RPU for the $100. Likewise, one should be able to substitute miles for money (maybe with some minimum amount in cash) at a rate of $0.025/mile. If there is a minimum, it's tempting to say elites have a lower cash threshold, so that a silver can pay for an entire domestic upgrade in miles, while a GM has to pay at least $100 in cash before using miles. However, that allows United to game the system. It might mean you need to be able to participate in the auction without disclosing who you are, at which point you get a token that lets you upgrade, and gives you a refund on cash based on whatever miles and instruments you use. I'm guessing that's too complicated, though, and would not fly.
update: Also, in the interest of transparency, it would be nice if United published yearly stats. E.g., this is the fraction of CPUs that cleared by elite level, etc.
So for transparency, what I would do is consolidate into one upgrade list available well in advance and make it public. To address the privacy issue (which frankly is not great even under the status quo), instead of people's initials it should be part of the reservation number of something.
Now to make the upgrades strategyproof, you want the same pricing available to everyone. So first of all, United should restore the ability of people to buy up to premium cabins without a change fee. Second of all, the price of TOD upgrades should be publicly available (i.e., you can get it on the flight status page without logging in). Moreover, the effect of TOD on PQMs and PQDs should be disclosed before purchase (and ideally they would all be some sort of buy-up that counts towards PQD).
Many strategyproof techniques involve some sort of auction. Thus, to give elites some advantage you should be able to choose how to pay for an upgrade after getting the upgrade. For example, if a TOD is available for $100, I should be able to buy it, then, if desired, substitute an RPU for the $100. Likewise, one should be able to substitute miles for money (maybe with some minimum amount in cash) at a rate of $0.025/mile. If there is a minimum, it's tempting to say elites have a lower cash threshold, so that a silver can pay for an entire domestic upgrade in miles, while a GM has to pay at least $100 in cash before using miles. However, that allows United to game the system. It might mean you need to be able to participate in the auction without disclosing who you are, at which point you get a token that lets you upgrade, and gives you a refund on cash based on whatever miles and instruments you use. I'm guessing that's too complicated, though, and would not fly.
update: Also, in the interest of transparency, it would be nice if United published yearly stats. E.g., this is the fraction of CPUs that cleared by elite level, etc.
Last edited by ftweb; Apr 10, 2016 at 9:14 am
Apr 10, 2016, 9:23 am
#37
Join Date: Jan 2013
Location: BOS
Programs: Hyatt Discoverist, Marriott/SPG/Hilton Gold, PreCheck + Clear
Posts: 2,306
Last edited by RandomBaritone; Apr 10, 2016 at 9:30 am
Apr 10, 2016, 9:36 am
#38
Join Date: Apr 2015
Programs: United Global Services, Amtrak Select Executive
Posts: 4,092
Revenue managers at UA are not stupid. They understand consumer behavior, and the fact is: Variable Ratio schedules of reinforcement lead to high rate behavior. That is exactly the type of schedule we are all on when it comes to receiving upgrades on UA. Higher status flyers may have a slightly different ratio schedule than lower status flyers, but the principle continues to operate to keep us salivating (apologies to Pavlov's dogs).
Apr 10, 2016, 10:48 am
#39
A FlyerTalk Posting Legend
Join Date: Apr 2004
Location: GVA (Greater Vancouver Area)
Programs: DREAD Gold; UA 1.035MM; Bonvoy Au-197; PCC Elite+; CCC Elite+; MSC C-12; CWC Au-197; WoH Dis
Posts: 52,122
Apr 10, 2016, 1:26 pm
#40
FlyerTalk Evangelist
Join Date: Feb 2002
Location: San Francisco/Tel Aviv/YYZ
Programs: CO 1K-MM
Posts: 10,762
More important than them not being stupid, they have access to an incredible vast quantity of empirical data concerning how UA flyers respond to various permutations and quantitative variations of the incentives. While we speculate here with little empirical basis about what effect X, Y, or Z incentive has on ticket purchasing behavior, UA actually *knows* what effects their incentive structures have. Of course, this is after the fact, and so when UA creates new incentive structures or varies them quantitatively, they are making predictions about what effects they will have. But once implemented, then learn very quickly.
I would like to see the elimination of both CPU and ToD. They can set mileage bands at something like $50/500 miles, and offer discounts on the upgrade fees by status, along with keeping the upgrade windows.
GS/1K = 50%, 1P(Gold/Plat) = 25%, 2P(Silver) 10%.
Everyone is "paying their way", and the status gets you priority and discounts.
I don't like the expectation that you'll get an upgrade if the delivery rate sucks. And I really don't like the idea that you could be #1 on the list as a 1K and some kettle checks in and pay $59 to "upfare" to F. In no case should a low/no status passenger get a better price/condition than a status passenger.
Apr 10, 2016, 1:35 pm
#41
Join Date: Mar 2005
Programs: fwp blood diamond, dykwia uranium
Posts: 7,251
Apr 10, 2016, 1:38 pm
#42
Join Date: Oct 2013
Location: ORD
Programs: UA Silver, Marriott Platinum/LT Platinum, Hilton Gold
Posts: 5,594
This is NOT the same market as even just 10 years ago where capacity was higher than demand on many routes. Planes are often full, with a standby list. The CPU system exists because you need some process to maximize revenue by not letting planes fly with empty seats.
So, if UA wants to monetize the heck out of F and sell 99% of F in some way, that's a business decision, although an unlikely one. But even then they need CPU's, or whatever name you want to give the process for filling those empty seats. The only other option is to fill the empty F seats with standby passengers, regardless of status. I'm guessing you don't want that though?
Apr 10, 2016, 3:20 pm
#43
Join Date: Dec 2012
Location: YEG
Posts: 269
No it was not intended as sarcasm, I do purchase the A and P fares when available. The organization I work for uses the lowest cost approach so I am on the hook for any flying at the front of the cabin.
Apr 10, 2016, 3:47 pm
#44
Suspended
Join Date: Mar 2002
Location: Canada, USA, Europe
Programs: UA 1K
Posts: 31,452
First principles:
1) On a per flight basis, United should try to maximize revenue by extracting the maximum amount of cash.
2) Generally, that means price discrimination -- discounting F cabin seats rather than letting them go empty while still defending the price of full fare F.
3) Even an optimal pricing strategy will lead to unsold F seats.
4) United should give away those seats rather than let them fly empty because the cost of a flight is relatively fixed, and the goodwill generated by an upgrade will exceed the variable cost (incremental food and beverage, etc.), particularly if that upgrade is targeted.
5) One main priority of a loyalty program should be to allocate unsold inventory in a way that maximizes the value of the goodwill generated by targeting the customers who will on average alter their future purchasing decisions the most as result of the upgrade.
6) [Controversial] United's actions lead me to believe that they do not believe that, on average, the goodwill generated even by upgrading an elite with an instrument is worth more than $x00. That may in fact be right or wrong, but it seems clear that that is the belief at least. Oscar's most recent comments may indicate some change in that thinking/analysis.
CONCLUSIONS/ASSERTIONS
1) CPU are largely a red herring.
2) The emphasis should be on instrument/mileage based upgrades.
3) This means that status should be only indirectly considered. It should matter only insomuch as people with higher status will also have more instruments and miles. Note that this only applies to upgrades -- other status perks might make sense.
PROPOSAL
1) The most theoretically efficient means would be an auction/bidding based system where anyone regardless of status can "bid" for upgrades on a waitlist basis. At some point before the flight when additional seats are unlikely to be sold (T-3ish?), the waitlist is cleared in order of people's bids.
In the interest of keeping the system manageable, you could imagine United allowing you to bid only at pre-set tiers, ie.:
any multiple of GPU
any multiple of RPU
any multiple of 10k miles
with some transparent conversion of GPU and RPU to miles and tiebreakers like time of check in or bid or something.
You could also imagine that with some data, UA could set some "buy it now" prices on a per flight that they are confident will be greater than the typical winning bid by a decent margin AND where the amount of work/margin etc that went into accumulating those instruments is enough to make it worth "forcing" or guaranteeing an upgrade.
After all, we can already guarantee a first class ticket with miles by booking one -- why should the ability to upgrade be any different? If I want to upgrade an econ ticket on PS from SFO to EWR, I can waitlist for R space for 30k miles or a GPU -- why not let me guarantee an upgrade for 60k miles (or some other number).
1) On a per flight basis, United should try to maximize revenue by extracting the maximum amount of cash.
2) Generally, that means price discrimination -- discounting F cabin seats rather than letting them go empty while still defending the price of full fare F.
3) Even an optimal pricing strategy will lead to unsold F seats.
4) United should give away those seats rather than let them fly empty because the cost of a flight is relatively fixed, and the goodwill generated by an upgrade will exceed the variable cost (incremental food and beverage, etc.), particularly if that upgrade is targeted.
5) One main priority of a loyalty program should be to allocate unsold inventory in a way that maximizes the value of the goodwill generated by targeting the customers who will on average alter their future purchasing decisions the most as result of the upgrade.
6) [Controversial] United's actions lead me to believe that they do not believe that, on average, the goodwill generated even by upgrading an elite with an instrument is worth more than $x00. That may in fact be right or wrong, but it seems clear that that is the belief at least. Oscar's most recent comments may indicate some change in that thinking/analysis.
CONCLUSIONS/ASSERTIONS
1) CPU are largely a red herring.
2) The emphasis should be on instrument/mileage based upgrades.
3) This means that status should be only indirectly considered. It should matter only insomuch as people with higher status will also have more instruments and miles. Note that this only applies to upgrades -- other status perks might make sense.
PROPOSAL
1) The most theoretically efficient means would be an auction/bidding based system where anyone regardless of status can "bid" for upgrades on a waitlist basis. At some point before the flight when additional seats are unlikely to be sold (T-3ish?), the waitlist is cleared in order of people's bids.
In the interest of keeping the system manageable, you could imagine United allowing you to bid only at pre-set tiers, ie.:
any multiple of GPU
any multiple of RPU
any multiple of 10k miles
with some transparent conversion of GPU and RPU to miles and tiebreakers like time of check in or bid or something.
You could also imagine that with some data, UA could set some "buy it now" prices on a per flight that they are confident will be greater than the typical winning bid by a decent margin AND where the amount of work/margin etc that went into accumulating those instruments is enough to make it worth "forcing" or guaranteeing an upgrade.
After all, we can already guarantee a first class ticket with miles by booking one -- why should the ability to upgrade be any different? If I want to upgrade an econ ticket on PS from SFO to EWR, I can waitlist for R space for 30k miles or a GPU -- why not let me guarantee an upgrade for 60k miles (or some other number).
Apr 10, 2016, 3:49 pm
#45
Suspended
Join Date: Mar 2002
Location: Canada, USA, Europe
Programs: UA 1K
Posts: 31,452
If it were idiosyncratic to my (extremely large) employer, that would be one thing. But doesn't almost every large employer use a corporate travel system with a similar restriction?
Put differently - who, exactly, flies regularly on business but actually has the flexibility to book front cabin at issuance domestically?
This isn't just me - this is United's bread-and-butter customer base.
Put differently - who, exactly, flies regularly on business but actually has the flexibility to book front cabin at issuance domestically?
This isn't just me - this is United's bread-and-butter customer base.