Starbucks finally got its customers back — transactions up 4.2% — and its revenue fell anyway, because it sold 60% of China
Comparable sales rose 7.9 per cent in Starbucks' third quarter, the fourth straight positive quarter of Brian Niccol's turnaround, and more than half of it came from people rather than prices. Consolidated revenue still fell 1 per cent, because 7,991 Chinese stores left the accounts in March. An analysis of a company that fixed the hard problem and deliberately shrank the easy one.
Every figure in this article is drawn from Starbucks' third-quarter fiscal 2026 earnings release and Form 10-Q, the company's China joint-venture announcements, or from CNBC and Restaurant Dive as attributed. Starbucks' fiscal year ends in late September. Sections marked as analysis are identified as such.
There is one number in a restaurant company's results that cannot be manufactured, and Starbucks just posted it. In the quarter ended 28 June, global comparable sales rose 7.9 per cent — and 4.2 points of that came from transactions, not price. In North America, transactions rose 4.5 per cent.
That distinction is the whole difference between a turnaround and a price increase, and it is worth stating plainly because most of the restaurant industry spent this year failing to manage it. Starbucks did not raise its way to a good quarter. More people bought coffee.
The quarter
It was the fourth consecutive quarter of positive global comparable sales and the second consecutive quarter of operating-margin expansion under the "Back to Starbucks" plan Brian Niccol announced within weeks of arriving in 2024. Non-GAAP operating margin reached 14.4 per cent, up 430 basis points. Non-GAAP earnings of $0.85 a share rose 70 per cent from $0.50; on a reported basis, earnings of $0.91 rose 86 per cent.
Two large one-off items sit inside that reported figure and pull in opposite directions: a $536.3 million net gain on the divestiture of the China retail business, and $302.6 million of restructuring and impairment charges against just $20.8 million a year earlier. Strip both out and the non-GAAP number — up 70 per cent — is the one to trust.
Underneath the one-offs the operating business genuinely improved. Reported operating income rose to $980.4 million from $935.6 million and GAAP operating margin gained 60 basis points to 10.5 per cent — modest numbers that carry the $302.6 million of charges. Over the first three quarters of the year, revenue was still up 4.2 per cent at $28.8 billion. The company opened 175 net new stores in the quarter and finished with 41,304 worldwide.
| Q3 FY2026 comparable sales | Total | Transactions | Ticket |
|---|---|---|---|
| Global | +7.9% | +4.2% | +3.5% |
| North America | +8.1% | +4.5% | +3.5% |
| International | +5.7% | +2.6% | +3.1% |
And yet consolidated net revenue fell 1 per cent to $9.3 billion. A company whose stores sold 8 per cent more in North America reported less revenue than a year ago. The explanation is not in the coffee business at all.
What happened to China
On 30 March, Starbucks closed a transaction it had signalled the previous November: funds managed by Boyu Capital acquired 60 per cent of its China retail operations for about $4 billion, with Starbucks retaining 40 per cent and continuing to own and license the brand and intellectual property to the new joint venture. The company put the total value of what it holds and earns from China — the sale proceeds, the retained stake and future licensing fees — at more than $13 billion.
In accounting terms the consequence was immediate and mechanical. 7,991 company-operated stores stopped being Starbucks' stores and became licensed stores. Their full retail revenue left the consolidated top line and was replaced by a royalty. International segment revenue duly fell from $2.01 billion to $1.32 billion in a single year, while North America rose from $6.93 billion to $7.40 billion.
This is why revenue and margin moved in opposite directions. Licensing revenue is smaller and far more profitable than operating a coffeehouse: no baristas, no leases, no milk. A large part of the 430-basis-point margin expansion is simply the arithmetic of removing eight thousand shops from the income statement.
It also strips out the volatility. China had been Starbucks' most contested market for five years — a price war with local chains selling coffee for a third of Starbucks' ticket, a consumer economy that stopped cooperating, and a brand that no longer had the field to itself. Boyu now owns that fight. Starbucks owns a royalty on it.
Analysis: the number that could not be faked
Set Starbucks beside the rest of the sector and the achievement sharpens. Across American restaurants in 2026, comparable sales have been carried almost entirely by average check while guest counts fell; McDonald's posted 0.8 per cent US comparable-sales growth in its own second quarter on higher checks and negative guest counts. Starbucks is the conspicuous exception, and it got there by spending money rather than discounting.
The mechanism management credits is Green Apron Service — more labour hours in stores, standardised routines, coaching and accountability, plus the reversal of a decade of decisions that had turned the cafés into pickup counters. It is an operating fix, not a marketing one, and it is expensive in exactly the line item that Wall Street usually wants cut.
The digital side is doing its part rather than carrying the quarter. Starbucks counted 35.8 million 90-day active Rewards members in the United States — a base large enough that its behaviour is effectively the American speciality-coffee market, and one that management has spent the past year deliberately not spamming with discounts. The recovery has been built on service throughput and menu simplification, not on offers.
Our Back to Starbucks plan was built on the belief that an extraordinary cup of coffee, human connection and customer experience win the day, every day. Our third quarter results are proof they do. — Brian Niccol, chairman and chief executive
What it cost to get here
The restructuring charges are the receipt. Starbucks has taken $415.8 million of restructuring and impairments across the first three quarters of fiscal 2026 against $137.0 million in the comparable period, covering the reorganisation of its support functions and the closure of company-operated stores that did not fit the new format. The turnaround has been paid for in headcount and leases as well as labour hours.
Guidance now assumes the improvement holds: fourth-quarter US comparable sales of 6.5 per cent or better, full-year global comparable sales near 6 per cent, non-GAAP operating margin above 11 per cent and non-GAAP earnings of $2.55 to $2.65 a share, with 600 to 650 net new stores. Consolidated revenue for the year is guided to flat-to-slight growth — the China hole, again.
Chief financial officer Cathy Smith framed the year in the language of a company that no longer expects help from the macro environment: "We are focused on what we can control amid a dynamic operating environment — executing our Back to Starbucks plan with discipline." Two years ago Starbucks was explaining away declining comparable sales in both of its largest markets simultaneously. The bar has moved.
Analysis: a smaller company on purpose
Starbucks ended the quarter with 41,304 stores, two-thirds of them licensed rather than company-operated. That ratio has been drifting for years and the China deal moved it decisively. The direction of travel is toward the model McDonald's has run for decades: own the brand, license the operations, collect a royalty, keep the margin and hand the capital intensity to somebody else.
The North America business now supplies $7.4 billion of the $9.3 billion of quarterly revenue — close to 80 per cent — against $1.3 billion from everything international and $588 million from the packaged-coffee Channel Development arm. For a brand that spent two decades describing itself as a global growth story, that is a striking concentration, and it makes the American consumer the single point of failure in a way it has not been since the 1990s.
The trade is real in both directions. Starbucks has converted the most capital-hungry, most politically exposed part of its estate into $4 billion of cash, a 40 per cent stake and a royalty stream, and its reported margins look better for it. It has also capped its own upside in the market with the most coffeehouses left to build — the joint venture's stated ambition is as many as 20,000 Chinese locations, and Starbucks will now receive a licensing fee on them rather than the sales.
The judgement to make about this quarter is therefore not whether the turnaround is working. Transactions up 4.2 per cent settles that; you cannot fake people walking in. The judgement is about what Starbucks chose to be while fixing itself: a smaller, more profitable, more American company that has decided the fastest-growing coffee market on earth is better rented than owned.