#121: United made me do it: incentives and side effects

May 24, 2023; 16 min read

I have a long history of reading the New Yorker.

When I was little, we used to flip through the magazine for the cartoons; later, I came to appreciate the long format articles — which are well-edited, making them more readable and enjoyable. Over the years, my New Yorker subscription has come and gone; the buildup of weekly paper copies seems wasteful for a nomad, and the digital version is too easy to forget about.

The other day I was reading the header of a New Yorker online article that was behind their paywall, and I’d already read my allocation of free articles for the month. The subscription page had an offer of $6 for 12 weeks (which would automatically convert to $120/year), which is pretty much giving it away (especially if you consider the weekday New York Times in LA is $6.50 for just one issue on a newsstand). I considered the offer — but the deal was killed by the “incentive” — the tote bag.

Not only do I not want the tote bag, but like Groucho Marx, I’m not sure I want to be a member of any group where their decision to do something was based on a tote bag. I guess it’s a historical thing — I think I recall tote bags being part of the new subscriber offer for decades, but it’s time to update that incentive (or at least offer more modern alternatives).

In this issue I’ll tackle some of the issues around incentives, and side effects. And it gives me an opportunity to transition to one of my favorite topics: airline mileage programs, one of the most powerful behavioral modification systems ever invented.

The ADA effect

I can’t find any confirmation of this online but I once heard from a movie industry insider that the Americans with Disabilities Act (ADA) had had the opposite effect from its intention — because it required disabled accessibility modifications for all movie theaters over 100 seats (which was once the norm), but after the passage of the ADA in 1990, this helped accelerate a shift towards multiplex theaters where each individual theater was less than 100 seats and therefore no disabled access was legally required — reducing costs, increasing yield, and perversely reducing access for people with disabilities. The rise of the multiplex wasn’t just ADA-related, it was also related to efficient capacity utilization and variety, as well as the advent of digital distribution which reduced the hassle and expense of showing any one movie. Nevertheless, stories where incentives backfire are somehow quite satisfying. 

Chinese policies

In the mid-2000s, the Chinese economy was growing like wildfire, and real estate speculation was particularly explosive. In order to curb speculation, in some cities, they decided to restrict banks from offering aggressive interest rates for non-primary residence mortgages, which seems rational as a brake on speculation. When the real estate market, especially the high end segment, started to rise even faster, they were confused — but the investor logic was that if I can only get one attractive mortgage, then I should combine my primary and speculative budgets, going all-in on my primary residence… thereby accelerating the prices of high end real estate.   

Staying in China, there is a well-known Four Pests campaign about Chairman Mao initiating a program to kill sparrows and other pests to reduce their grain pilferage, which was “successful” but which had the side effect of letting the locust population run wild, thus decimating the grain crops far more than the sparrows ever did, leading to widespread famine. 

The one-child policy, in effect from 1979-2015, seemed like a reasonable response to rapid population growth, now has led to one of the most extreme demographic issues, both in terms of dependency issues (too few workers supporting too many retired dependents) as well as a highly skewed male female ratio because of the preference for male offspring (giving eligible women very strong bargaining positions).

The last China story I’ll offer is the one about the use of car horns in Shanghai — the constant use of horns became annoying to residents, so the authorities imposed penalties for using them (how did they intend to enforce this?); the result of the horn ban was that impatient drivers would tailgate at even closer (unsafe) distances and flash their lights, resulting in the desired effect of less horn noise, along with the undesired effect of more car crashes.

Short term behavior and long term effects

These missteps are due to the lack of understanding, either of consumer logic, or the  differentials between simple first order short term behavior and complex long term outcomes, which include secondary and tertiary effects.

Since incentives can influence the target audience in unintentional ways, it makes sense to incorporate feedback and make the programs adaptive rather than static, so that the linkages between the incentive system and the end objective can be constantly monitored and calibrated. 

Sometimes a system will seem fine as it is, and then a competing system will come along and materially change what used to be an equilibrium, like LIV Golf changing the competitive environment for the PGA, which had become comfortable in its power over the players and exhibited correspondingly high inertia as a result. Therefore, adaptive methods might be more productive. In the area of consumer behavior, digitization and AI make A/B testing easier and more effective, so various possibilities can be empirically tried and tested at smaller scales without having to make single large bets and merely hoping for the best.     

Another way to think about incentives is to look at programs which seem to work well. One of my favorite examples, partly because they definitely affect the behavior of millions, but also because I’m always trying to find and exploit their collective loopholes, are the airline loyalty programs. And one of their strongest features is that they are constantly adapting to changes in client preferences, as well as a dynamic competitive environment.

Airline loyalty programs

One of the most powerful behavioral strategies ever devised, airline loyalty programs essentially began in the early 1980s; there are trillions of miles outstanding on all of the major airlines, probably worth well over $200 bn – the latest data I could find was an Economist article suggesting a total of 14 trillion miles back in 2005, suggesting many multiples of these amounts outstanding in 2023. Although I am wary of overweighting my own data, just within the US Big Three, 86% of my flights have been on United in the last 7 years, compared to only 11% for Delta, and 2% for American, because I’m heavily invested in the United MileagePlus program and have zero status at the other two.

Once affinity is established, customers favor that airline when alternatives are available; they spend some time to understand the rules to gain status, which modifies their behavior — and the programs provide them with sufficient value in terms of flights and perks to make maintaining or raising their status a priority (within reason). 

A large chunk of airline miles are “earned” as a corporate expense — paid for by a corporation, but with mileage accruing to the *individual*, which is kind of a sneaky way to give a little kickback to the individual. In that sense, those miles do seem “free” because the individual is getting a benefit from an activity they didn’t pay for (especially if the flight was a business class ticket which earned lots of miles and also helped their pursuit of status). In addition, many corporate credit cards are a source of miles which accrue to the individual rather than to the corporate.

The perks of loyalty

From the client point of view, the primary benefits of loyalty programs include mileage accumulation for ‘free’ flights and upgrades; reaching certain status levels can result in priority boarding, lounge access, and a variety of other VIP-like perks. Especially for free flights and upgrades, the prospect of not paying for what are usually significant expenses is quite compelling.

“Free” flights

One of the central “tricks” of loyalty programs is a version of bait-and-switch: they want clients to believe they can use miles for anything when in fact it’s only the dregs that appear as great deals, that only fliers with great flexibility can access; high value peak period tickets require far more miles. This is why they offer a plethora of other perks which cost relatively little to provide, in order to give customers the sense that their loyalty is “valued” even if they usually can’t get the flights they want.

The simple “buy ten, get one free” loyalty programs at your local coffee house is just a 10% discount for repeat customers; the first coffee and the 10th (free) coffee are essentially the same quality, because there is no volatility in the price. The rough equivalent of the airline loyalty program would be “buy ten fresh coffees, get one day-old coffee free” in the sense that the “free” items are NOT the same as the other items purchased, they are materially less valuable to the airline. 

There are many sites like this one dedicated to showing people the best ways to use these mileage programs; for the most part they agree that most airline miles are worth about 1-2 cents each. Although the original loyalty programs generated one mile per mile flown, this equation has changed dramatically so now there is only a loose connection between mileage earned and mileage flown, but a reasonable connection to dollars spent.

There are a variety of sites (ThriftyTraveler, SeatSpy, et al) which purport to help travelers use miles to navigate the mileage program ecosystem to score inexpensive flights — my airline geek friends and I collectively call this the “black arts” — which is one topic I’ll never write about, because not all secrets needs to be shared (though the sites above are a step in the right direction). Most of these sites are sponsored by credit cards hoping they will entice you to open an account, so despite the seemingly attractive mileage offers, it’s clearly a positive sum outcome for them too. 

Desperately seeking status

Loyalty programs entice frequent travelers to travel on their airline and spend more to attain higher tiers of status.  In addition to upgrades and ‘free’ seats, status comes with the incentives of priority boarding, dedicated check-in lines, lounge access, luggage allowances, preferential upgrades — and even complimentary limo transfers —  as Sirens seducing you to the sweet pleasures of higher and higher status. 

Outside of the tangible benefits like dedicated check-in lines, I wonder why status seems to matter so much — I fully admit to being enamored of being seen as an experienced frequent flier, with a fat passport filled with exotic stamps, passing through security lines and immigration lines as if I was covered in Teflon, even if most of those hacks have nothing to do with airline status.

Originally, status was directly correlated to aggregate mileage flown. Since fare prices vary so much, airlines now largely focus on actual dollars spent rather than miles flown — although the miles flown are the same, business class fares are multiples of economy class fares, so status is driven by dollars spent, and the metric of miles flown is increasingly irrelevant from the status perspective.

As a perk, I am routinely surprised at how effective priority boarding is — given the fact that the plane won’t leave until everyone is on board, boarding order should make relatively little difference – although the advent of paid checked baggage means a bias to carry on luggage, with the result that overhead bins are almost always overflowing. But since airlines make the process very transparent, they are adding the additional dimension of social status to the mix (who gets to put their rollerboards into empty bins, and who gets to stress about finding room).

On United, for instance, the top tier invite-only Global Services members get to pre-board a few seconds earlier than 1K members, who board just before group 1; I find myself inordinately focused on this silly metric, which is probably related to where you explicitly stand in the United pecking order.  Once when I was flying Delta, where I have zero status, and I was in boarding group 9 (!); I’d never heard of such a lowly group, and thought I might be in with the pets in the baggage compartment. There are situations where someone who doesn’t *look* like a high status customer will attempt to priority board; those who have the status are sometimes offended that they have been questioned; those who are not entitled are embarrassed by being  turned back at the gate. 

Obviously there has to be *some* boarding order; the fact that airlines can indirectly monetize this through their loyalty programs is a kind of alchemy.  Some studies suggest boarding window seats first is optimal for efficiency (I’ve done this on a discount airline) — it’s no surprise that boarding fat cats in the front of the plane first is better for making them feel special than boarding efficiency.

Revenue management

Airlines have “revenue management departments” which aim to maximize their efficiency by getting the most per seat possible, and filling as many seats as they can. 

It comes as no surprise that the cost of fulfilling these points is far less than their purported value, because the airline business is not like the grocery business, for instance, which has relatively constant demand; airline demand is highly cyclical, with lots of peaks and valleys. 

Most commercial airline flights have a daily frequency and therefore supply is relatively static, but demand patterns are cyclical. Business travelers prefer to travel during the beginning or end of the day to optimize their working schedule, for instance. Like most economic goods, seat pricing is a function of supply and demand, there are significant price differences depending on the day and time, and between non-stops and connecting itineraries.

Airlines decide what seats or upgrades can be accessed for “free” tickets, and how much they cost in miles, they can ensure they are “selling” those seats for more than they expect to get in cash. The headline teaser offers (“Transatlantic for only 60,000 miles in business class!”) which seem like great deals have major caveats about availability, eligibility, and surcharges.  The high demand seats during peak periods require significant redemption miles – often they cost the same or even more (at 2c/mile) than they would if you paid cash (so the airline sometimes gets a mark-up on an already fully-marked up price) — I’ve noticed Delta is particularly egregious about this, although increasing markups is also the general trend.

For the US based airlines, domestic lounge access is also an area where airlines have induced significant devaluation — sold to weary customers as oases of peace and luxury away from the unwashed masses, the modern US domestic lounge has become a crowded melee of frustration rather than a Zen paradise. Outside the US, however, the lounges of carriers like Emirates, Singapore, and Turkish Airlines routinely meet or exceed the promise of a luxury oasis. 

Adaptive challenges: devaluation and competition

The bad news about loyalty mileage is that mileage is routinely devalued — changing the award tables or the recent fully dynamic pricing initiatives — because there are no regulations regarding point conversions, the only side effects are potential reputational damage (so devaluation is done slowly but surely — like inflation at the supermarket which comes in downsizing package sizes rather than increasing price).

The reality is that compared to the 1.5-2c of “value” per mile that you are getting, it costs the airline far less than that, because they control what is available, how much mileage is required for a “free” ticket, and they routinely devalue your mileage inventory. Effectively they’re making a large gross margin selling miles (much of which they sell to credit cards) — precisely because they are so effective behaviorally, they have proliferated and penetrated into many areas of consumer spending. Since the nuts-and-bolts airline business is notoriously one of high cyclicality and low long term profitability, it’s not surprising that airlines love their loyalty programs. 

The good news is that because loyalty programs work, there is a lot of competition to attract customers, with the ability to earn miles rising rapidly. For airlines themselves, there is a disconnect between miles earned and miles flown — nowadays one can earn multiples of what one flies depending on status — on my last economy class flight, for instance, I earned miles equivalent to a 20% discount on the fare.  For credit card spending, the old standard of one mile per dollar spent has risen to two, with special use cases getting 5 or more.

Therefore the longer term trend will probably be a continued increase in the ability to earn miles, which will be made very explicit in their ads; and steady devaluations of those same miles, which will be done as quietly as possible. It suggests an optimal consumer strategy of keeping miles at the credit card level (where they are less subject to devaluation), and spending airline mile balances as soon as possible (on non-peak flights, if possible) before they are devalued or expire. 

The side effects of loyalty

If loyalty programs work so well, then it makes sense to ask if there are detrimental side effects. I think it makes sense to separate earning status (largely via flying) vs earning miles (both flying and spending). If you fly frequently, it should be clear that status is harder to earn, and probably more valuable to have, than mileage.

On the mileage side, since miles can be earned more readily on credit cards than airline tickets, this could act as a further incentive to overspend on the credit card, because you feel you’re getting constant discounts. But since mileage benefits are being constantly devalued, this shouldn’t be an primary catalyst, although the signup bonuses for new credit cards can be sufficient incentive.

At the end of the year, some diehard frequent fliers used to do mileage runs (trips to random places simply to earn mileage) but since airlines have moved to revenue based status levels rather than mileage based levels, mileage runs now make less sense. The only side effect I can see is that business travelers might fly more often on the company budget to earn status, but on the whole I don’t see an enormous pattern of misbehavior due to a desire to win status.

In the hotel loyalty space, sometimes the program of the senior person you are traveling with will determine which hotel the group will choose; this is a minor inconvenience unless it happens all the time.

Loyalty program switching costs are notoriously high, to the point where many airlines will match status for limited times, to lower switching costs for frequent fliers of a different airline. 

Travel site recommendations

Here are a few sites I find useful for managing the flight ecosystem:

• my.flightradar24.com — track your flights and aggregate statistics; also displays a global map of your travel by year; very useful during tax time to determine time spent in different places

• expertflyer.com — find award seat availability; primitive user interface, however

• awardwallet.com — keeps track of your airline and credit card miles

Final thoughts

• Incentives often come with unintended side effects, so feedback and dynamic adjustment should be a part of the process

• Airline loyalty programs are quite effective in influencing consumer behavior, and have expanded into a wide range of other businesses

• Airline mileage is being constantly devalued because it increases airline profits; but mileage incentives also keep growing because the programs work and there’s competition to attract customers

• Tote bags represent a negative incentive for me