The Loyalty Program That Grew Every Number Except Loyalty

Airline loyalty programs have never had more members. They have also never been trusted less by the members they have.

Delta runs one of the largest loyalty programs in the world. Its members can earn points, book reward flights, and get lounge access. But the points are worth less than they were a few years ago, and the program keeps changing in ways that make it harder for customers to use them.

In September 2024, the U.S. Department of Transportation opened a formal inquiry into the loyalty programs run by American, Delta, Southwest, and United. The DOT's stated concerns were specific: devaluation of already-earned rewards, a shift toward dynamic pricing that makes redemption costs unpredictable, hidden or added fees, and reduced competition in the market for travel rewards.

The backdrop makes the inquiry easy to understand. Airline miles have lost roughly two thirds of their purchasing power since 2014, according to loyalty industry analysis. Delta's program alone shed close to a fifth of its per-mile value in the two years leading into 2026. Award charts that once reliably converted a fixed number of miles into a specific seat have been replaced, program by program, with dynamic pricing that moves against the traveler more often than it moves in their favor.

At the same time, elite qualification criteria shifted away from how often or how far someone actually flew, and toward how much they spent. Distance and segment counts, the metrics that used to reward genuine frequent flying, were quietly retired in favor of revenue thresholds that favor whoever puts the most dollars through a co-branded credit card.

None of this stopped membership rosters from growing. Airlines continued to report record enrollment. That is not a contradiction. It is the entire mechanism.

Delta is the clearest example, but the pattern runs across the industry, and the direction of travel is the same everywhere it shows up. Every major U.S. carrier has moved in the same three directions over the same period: less predictable redemption value, spend-based rather than distance-based status, and continued enrollment growth despite both.

The devaluation isn't even contained to the airlines that issue the miles. In January 2026, Capital One cut its transfer ratio to Emirates from a full one-to-one exchange down to 1,000 points for every 750 miles, a reminder that the banks sitting behind these programs are adjusting the same lever for the same reason. When multiple airlines and the card issuers behind them all move the same direction independently, that is no longer a company-specific pricing decision. It is a structural feature of how the loyalty economy currently works.

That is the detail worth sitting with before diagnosing the mechanism. A single airline devaluing its program is a pricing story. An entire industry devaluing in lockstep while enrollment keeps climbing is a behavioral story, and it is the kind regulators eventually notice, which is exactly what happened here.

Start with why these programs were so engaging in the first place, because the mechanism that built the loyalty is the same one now being exploited to erode it.

Variable ratio reinforcement: A reward schedule where a behavior is reinforced after an unpredictable number of repetitions rather than a fixed one. Behavior maintained on a variable schedule is far more resistant to extinction than behavior maintained on a fixed one, which is why slot machines and, structurally, loyalty programs, are built around unpredictable payout rather than guaranteed payout.

A loyalty program was never actually selling a fixed exchange rate. It was selling the occasional outsized win, a business class seat redeemed for a fraction of its cash price, that felt disproportionate to the miles spent earning it. That variability is what made the behavior sticky. It is also what makes members slow to notice, and slower to act on, a steady erosion in average value. The occasional win still lands. It simply lands less often and buys less when it does, and a reward schedule built on unpredictability is, by design, hard for the person inside it to audit in real time.

That explains why members tolerate devaluation longer than a rational actor would. It does not explain why airlines chose to devalue in the first place. For that, look at who the loyalty division is actually accountable to.

Principal-agent problem: A situation where decision-making is delegated to an agent whose interests diverge from the party they are meant to serve, so decisions optimize for the agent's incentives while still being presented as serving the principal.

Airline loyalty programs generate a large share of their profit from selling miles to co-branded credit card partners, not from the flights members eventually redeem. The member is the nominal principal the program claims to serve. The card issuer, and the fee income that relationship produces, functions as the real one. Enrollment growth directly increases card originations and fee revenue. Redemption value is a cost line. A loyalty division facing that incentive structure will rationally optimize for the metric that pays its bills, membership count, over the metric that costs it money, redemption value, even while marketing the program as a rewards system built around the member.

This is why the two numbers can move in opposite directions for years without the internal team treating it as a contradiction. From inside the business model, it isn't one.

There is a second-order effect that a pure revenue analysis misses, and it concerns what a devaluation signals to the members who are paying closest attention, not just the ones directly affected by any single change.

Elite and near-elite members are disproportionately business travelers, and business travel decisions are rarely made by the traveler alone. A frequent flyer who watches their status requirements shift from miles flown to dollars spent, while their existing balance quietly loses value, receives a clear signal about what the airline actually rewards. That signal does not stay contained to loyalty math. It shapes which airline that traveler recommends when their company negotiates a corporate travel policy, and corporate contracts, not individual bookings, are where an airline's most reliable revenue actually lives.

A loyalty program devaluation is rarely just a loyalty story. It is a trust signal broadcast to the exact population with the most influence over an airline's highest-value accounts, and that population is unusually well informed, unusually vocal online, and unusually able to act on what they conclude.

There is a reason this ended up in front of a federal regulator instead of resolving itself through ordinary market pressure, and it comes back to a concept that showed up in the first issue of this newsletter, applied here to a very different industry.

Members cannot independently verify whether a devaluation was necessary, fair, or applied consistently. Award charts disappear behind dynamic pricing algorithms that are opaque by design. Elite thresholds change with limited notice and little disclosed rationale. When a process affecting something members have already earned is neither transparent nor consistent, the perception of fairness collapses regardless of whether the underlying business rationale was reasonable.

That collapse is precisely what invites outside intervention. The DOT's inquiry exists because the market's usual correction mechanism, dissatisfied customers taking their business elsewhere, does not work well against variable ratio reinforcement and industry-wide lockstep devaluation. When private trust fails and private exit is structurally blocked, the next actor to step in is a regulator, and that shift changes the entire nature of the problem an airline is now managing.

The fix has to separate two things these programs currently blend together: acquisition and value delivery. Both matter, but tracking only the first while quietly degrading the second is the mechanism this entire issue has been describing.

In practice, that means:

• Report engagement, not just enrollment. Membership count should never be the headline metric on its own. Redemption rate, average redemption value relative to cash price, and repeat redemption behavior are the numbers that actually indicate whether the program is delivering what it promises, and none of them can be gamed by a co-branded card sign-up bonus.

• Restore some predictability to the reward schedule. A program does not need to abandon dynamic pricing entirely to rebuild trust, but a published floor, a guaranteed minimum award availability, or a fixed chart for a defined portion of inventory, gives members something stable to plan around rather than a constantly shifting target.

• Separate qualification from pure spend. A hybrid model that still credits actual flying, not only dollars charged to a card, keeps the program connected to the behavior it was originally designed to reward, rather than becoming indistinguishable from a general cash-back product wearing an airline's logo.

• Give real notice before devaluing earned balances. Grandfathering existing balances, or providing a genuine window to redeem at the old rate before a change takes effect, is the single clearest signal a program can send that it still considers the member the actual principal, not just the credit card partner behind them.

The fix is not a bigger sign-up bonus. It is measuring, and being willing to report, the number that has been allowed to quietly diverge from the one on the press release.

• The incentive to prioritize enrollment is not a mistake. It is the business model. Card partnership revenue is real, substantial, and unlikely to be voluntarily deprioritized just because it creates a principal-agent conflict. Any fix that assumes airlines will choose member value over the revenue line that currently outperforms it is assuming away the actual constraint.

• Predictability has a genuine cost airlines can point to. Dynamic pricing exists partly because fixed award charts leave money on the table during high-demand periods. A return to full predictability is not free, and airlines have a legitimate revenue management argument, not just a member-hostile one, for resisting it.

• Regulatory pressure moves slowly and does not guarantee structural change. A DOT inquiry can produce disclosure requirements without touching the underlying incentive structure that produced the problem. It is entirely possible for this story to end with more fine print and the same principal-agent conflict fully intact.

• This diagnosis assumes the devaluation was a byproduct of misaligned incentives, not the deliberate goal. If an airline's actual objective was extracting maximum value from a captive, sunk-cost member base before switching costs eventually erode, no amount of better reporting changes that outcome. This framework works best on programs that lost track of member value while chasing a different number. It has less to offer if the erosion was the plan.

If your organization runs anything with the word loyalty in it, genuine question: when was the last time someone on your team could tell you, without pulling a special report, whether your members are more engaged this year than last, separate from whether there are simply more of them?