In March 2026, Starbucks “reimagined” its Rewards program. Three new tiers — Green, Gold, Reserve. New earn rates. New ways to redeem. The press release used the words “more meaningful value, personalization, and engagement” in a single sentence, which is usually a sign that none of those things are about to happen.

Read the fine print and a different program emerges. To reach Reserve, the top tier, a customer needs 2,500 Stars in a year. At the base earn rate of one star per dollar, that’s $2,500 of Starbucks spend. About $48 a week. Roughly seven drinks. Every week. For twelve months. Just to unlock the multiplier that lets you earn rewards faster.

This isn’t a loyalty program anymore. It’s a toll booth. And the toll just went up.

What “more meaningful value” actually means

Strip the marketing language and the structure is simple. Default members earn 1 star per dollar. Spend $500 a year — about $10 a week, the price of a daily latte habit — and you become Gold, earning 1.2 stars per dollar. Spend five times that to become Reserve, earning 1.7 stars per dollar.

The earn rate gap between Green and Reserve is 70%. The spend gap is 5,000%. CX Dive’s coverage noted the obvious: the new structure rewards the customers who already spend the most, and effectively raises the threshold for everyone else.

Reserve tier requires roughly seven drinks a week, every week, for a year — to unlock a 70% earning bonus.

The 60-Star “$2 off any item” redemption that Starbucks is highlighting as a member win? At base earn rate, that costs $60 of spend to unlock $2 of value. A 3.3% effective discount. There are open-loop credit cards that beat this on every transaction, with no app and no tier maintenance.

The toll booth pattern

Loyalty programs follow a predictable arc. Phase one: launch generously, train customers to use the app, build habit. Phase two: start measuring program ROI as a cost center. Phase three: quietly increase the work required to earn a reward while celebrating “new flexibility” and “expanded ways to engage.” Phase four: customers stop noticing the rewards and start noticing the friction.

Starbucks has been running this play for years. The 2023 changes nearly doubled the stars required for popular redemptions like brewed coffee and breakfast sandwiches. The 2024 mobile-order surcharge tests pushed cost back onto customers without touching the menu price. The 2026 tier launch is the next move in the same sequence: extract more spend per reward, frame it as personalization, hope nobody runs the math.

Here’s the part that should make every loyalty leader pause. Starbucks Rewards isn’t broken. By the metrics most CMOs report — active members, app downloads, reward-attached transactions — it’s a success story. 34 million active members in the US. The program drives a meaningful share of total transactions. From the deck, this is what winning looks like.

From the customer’s seat, it’s a slow leak. And once a customer starts noticing the leak, they don’t usually file a complaint. They just stop opening the app.

The choice that isn’t a choice

The deeper problem with the new tier structure isn’t the math. It’s what it asks the customer to do — or rather, what it doesn’t ask. Real engagement design gives customers interesting choices: decisions that feel meaningful, surface preference, and reward consequence with consequence.

What does the new Starbucks program ask the Reserve customer to choose? Nothing. Drink the coffee you were going to drink anyway. Open the app. Tap to claim. The “reward” for a year of Reserve-level spend is a marginally better extrinsic bribe on the next transaction. There is no narrative, no progression curve worth engaging with, no decision the customer makes that changes what they get. It’s not a game. It’s a meter.

Compare this to what gamification — real gamification, not the term as it’s used in loyalty decks — actually does. A well-designed engagement loop gives the player escalating choices with real consequence: which path to take, which resource to spend, which risk to accept. Each decision reveals something about what the player wants, and the system responds. The loop tightens. Engagement compounds. That’s the design philosophy behind PUG’s Picnic platform — turning passive customers into active participants by surrounding them with choices that feel like theirs to make.

Starbucks has none of this. It has a punch card with a CRM bolted on.

Why it still “works” — and why that’s the real risk

The defense from any loyalty leader reading this will be the same: “It’s working. The numbers are up.” And they are. Repeat purchase from rewards members beats non-members. The app drives mobile order volume. Members spend more per visit. All true.

All also true of every loyalty program in its phase three. The metrics that confirm the program is “working” are exactly the ones that don’t pick up disengagement until it’s too late. Active members stays high until it doesn’t. Reward-attached transactions look strong until customers stop bothering. By the time the dashboard reflects the rot, the rot is already a quarter old.

The signal to watch isn’t average transactions per member. It’s something quieter: the percentage of members who open the app without making a purchase. The drop in members who push past the home screen. The slow flattening of redemption rates across the middle of the customer base. None of these show up in a board deck. All of them are leading indicators that a program has crossed the line from engagement engine to transaction tax.

What loyalty leaders should take from this

Every program eventually faces the choice Starbucks just made. The CFO wants the breakage. The marketing team wants the tier story. The engineering team wants fewer exceptions. And the customer — somewhere down the priority list — wants to feel like the relationship is mutual.

A few things worth holding on to:

1. If your “premium tier” requires the customer to spend like it’s a full-time job, it’s not a tier — it’s a tax. Reserve at $2,500 a year isn’t aspirational. It’s a description of an existing customer the brand wants to lock in harder.
2. “Personalization” is not a synonym for “harder to earn rewards.” If the only thing personal about your program is the offer the algorithm thinks will get one more transaction out of you, customers will see through it inside a quarter.
3. Watch the friction, not the activity. The metric that matters is how often customers choose to engage when they don’t have to. Every other number is a lagging shadow of that one.
4. Phase three is a choice, not a destination. Programs don’t have to age into toll booths. They do because the org chart rewards the team that hits this quarter’s number.

Starbucks Rewards in 2010 was one of the most-copied programs in the industry. Every retailer benchmarked the punch-card-on-an-app model. In 2026, it’s worth asking what current Starbucks is being copied — and whether the people copying it have noticed that the original is quietly running out of road.

The best loyalty programs make customers feel chosen. The worst ones charge them for the privilege of staying in line.