Starbucks Rewards: A Loyalty Program Case Study

Starbucks Rewards is the most studied loyalty programme in the world, and for most of its life the lesson people took from it was the wrong one. The lesson was never the Star. It was that the programme and the ordering experience were the same thing, which turned loyalty from something customers consulted occasionally into something they used every time they bought coffee.

In March 2026 Starbucks rebuilt the programme around three status tiers for the first time since 2019, and in doing so it provided a second, less comfortable lesson about what happens when a programme changes the maths on people who have been earning under the old one.

This case study covers how the programme works now, what made it successful, what the 2026 overhaul changed and why it provoked a backlash, and what is actually transferable to a programme that is not Starbucks.

How Starbucks Rewards works

Programme terms change regularly, and the detail below describes the structure as at the time of writing. Check the official terms before relying on any specific figure.

The three tiers

At launch, existing members were placed into tiers based on their 2025 activity: roughly the top 5% began at Reserve, the next 20% at Gold, and everyone else at Green. That matters, because it means the structure was calibrated to the existing base rather than aspirational.

The redemption ladder

Redemption runs from a small customisation at the bottom of the ladder up through drinks, food and merchandise, with a 60-Star tier added in the 2026 revamp giving a fixed amount off any item. The low first rung is the most copied feature of the programme and the most important one: it means almost every member reaches a redemption rather than accumulating a balance they never use.

The app as the programme

Ordering, payment, earning, redemption and personalised offers all live in one place. A member does not open the app to check a balance; they open it to buy coffee and the balance is there. This collapses the engagement problem that almost every other programme spends its budget on.

What made the program successful

Six factors, and the first one explains more than the other five combined. It is also the one most often missed by brands trying to copy the programme.

The programme is the transaction

The single most transferable insight. Where loyalty sits inside the purchase flow rather than alongside it, engagement is not a separate thing to be bought with messaging. Most programmes that try to copy Starbucks copy the Star and not this, which is why they do not work.

Stored value changes the economics

Preloading funds onto a card or app balance pre-commits spending and smooths cash flow in a way no points programme does. It also creates a switching cost that has nothing to do with points: money already loaded is money already spent with you.

A reachable first reward

The bottom of the redemption ladder has always been low enough that an ordinary customer gets something quickly. Time to first redemption is the strongest predictor of whether an enrolment becomes a habit, and Starbucks has consistently optimised for it.

Personalisation with enough data to be relevant

Because the app is the transaction, the programme sees what was bought, when, where and how often. Personalised challenges and offers built on that data are relevant in a way that offers built on segment membership are not.

Gamified challenges rather than pure accrual

Bonus Star challenges, streaks and limited-time missions shape behaviour: try a category, visit on a slow day, come back within a window. This is the difference between rewarding what customers already do and changing what they do.

Operational consistency underneath

None of the above would work if the coffee were inconsistent or the queue unmanageable. The programme amplifies a reliable experience; it does not create one.

The 2026 overhaul and the backlash

The redesign is as instructive as the original programme, for different reasons. It is a live demonstration of what happens when a programme changes the maths on people who have been earning under the old one.

What changed

• Status tiers returned: Three tiers replacing a single flat structure, with earn rate increasing by tier.
• Earning became differentiated: Higher-spending members now earn Stars faster, which is the point of tiering.
• Expiry was split by tier: Entry-tier Stars lapse after around six months; upper-tier Stars do not expire.
• A lower redemption rung was added: A 60-Star option for money off any item.
• Some perks were capped: Previously open-ended benefits were limited, which drew separate criticism.

Why it provoked a backlash

Because for the majority of members the change was a devaluation, whatever else it was. Analyses circulating after launch estimated that many members would earn materially fewer Stars than before, and those estimates are contested, but the perception was immediate and widely shared. Expiry at the entry tier made it sharper: the customers least able to reach the next tier were the ones whose Stars now lapsed.

The lesson in the backlash

Devaluations are the most complained-about event in loyalty, and the complaint is rarely about the new structure in isolation. It is about the change applied to value people believe they already earned. Three things reduce the damage: long notice, grandfathering accumulated value, and making the first visible effect of the change a gain rather than a loss. Starbucks did the third, with the new low redemption rung, and still took the criticism, which indicates how little margin for error exists here.

What is actually transferable

Starbucks has advantages most businesses do not: daily purchase frequency, a low ticket, a universally installed app and enormous scale. Copying the mechanics without the frequency produces a programme that cannot work. These five things do transfer.

1. Put the programme inside the transaction: Whatever your purchase flow is, that is where the loyalty should live. A separate destination is a tax on participation.
2. Make the first reward reachable quickly: Low first rung, small reward, early. This is the single highest-return design decision available.
3. Use challenges to shape behaviour: Reward the visit you want rather than the spend you already have.
4. Consider stored value if frequency supports it: Pre-commitment changes behaviour and cash flow more reliably than points do.
5. Treat any devaluation as a crisis: Notice, grandfathering and a visible gain alongside the loss, or expect the response Starbucks received.

What does not transfer

• Daily frequency: Challenges and streaks need purchase occasions. In a category bought four times a year, they are irrelevant.
• A universally installed app: Most brands will never justify an app, and a wallet pass is the honest alternative. Our guide to loyalty mobile apps covers that decision.
• Scale economics: Programme costs that work across tens of millions of members do not work across tens of thousands.
• Tolerance for tiering: Tiers only motivate when the next tier is visible from where an ordinary customer stands.

What to measure if you are copying it

Building a comparable program on Loyalife

You can also explore restaurant loyalty programs and the Loyalife platform.

The Starbucks lesson has always been structural rather than mechanical. Put the programme where the purchase happens, make the first reward easy to reach, and understand that the moment you change the maths, every member recalculates what you are worth to them.