Starbucks: Two CEOs, two very different gameplans

One quick thing before the report. Today we're opening the doors on Nerd Out.

Years ago I wanted to build software and hit the same wall everybody hits. Learn to code, or write a big check to somebody who already had. I got Harvest Profit over that wall through pure stubbornness, and it eventually became part of John Deere. The wall is a lot shorter now.

What surprised me was that the exit wasn't the most rewarding part. It was a big win for our employees, for me, and for my family. But the real joy was in how work felt while I was building. I looked forward to every day. Real problems I cared about, for customers I loved working for. That feeling is the whole reason Nerd Out exists.

Most AI demos end with a list of things to try later. Ours end with something useful on your screen. You bring one part of your work that needs to be better: a repetitive task, some back-office mess, the first piece of an idea you keep circling. We spend 90 minutes building a working first version together and narrating what we're doing as we go, so you can keep improving it after we hang up. No coding required.

Our first workshop is live today, on sale tomorrow, and will be hosted on Thursday at 10 AM CT: NerdOut.com

Today, I'm digging into Starbucks Corporation (SBUX).

They sell coffee out of roughly 41,000 stores worldwide. Beverages are about 61% of revenue, food is 19%, and the rest is packaged coffee, licensing, and merchandise.

They came up as I'd been focusing on companies that have negative working capital. I knew they were going through some issues, as they recently brought in a new CEO. But it's been interesting to dig into Starbucks and learn more about their stumbles.

A few things that stood out to me:

• Operating profit fell 46% in FY2025 while revenue went up. They voluntarily gave back $2.5B of annual profit to put labor, hours, and training back into stores.

• Store operating expenses went from $14.7B in FY2023 to $17.1B in FY2025 on about $1.2B of additional revenue.

• Their investments appear to be paying off. Q2 FY2026 global comps were up 6.2%, with 3.8 points of transaction growth on top of 2.3 points of ticket/price.

• Regarding negative working capital, customers hand Starbucks about $1.8B before they drink anything. Gift card and app balances sit on the balance sheet as a liability called stored value, and due to it working capital ran negative $2.8B in FY2025.

• We have quite a bit of debt. Their net debt is $22.9B at the end of 2025 which was 5.6x their operating cash flow.

• They have ongoing labor issues. Union baristas struck for 131 days before bargaining resumed in April 2026. Labor is both the fix and the fight.

This is one of the more fun reports I've written, not because of my familiarity with Starbucks, but because of the CEO turnover and their different approaches to improving the company’s topline performance.

Laxman Narasimhan ran the company from March 2023 until the board pushed him out in August 2024, and Brian Niccol took over that September after six years turning Chipotle around.

Both guys were focused on fixing declining same-store sales, but they reached for different tools.

Narasimhan treated it as a demand problem. Give people a reason to come back:

• Opened the app to non-Rewards customers

• App-only offers that flexed by location, inventory, and weather

• Meal deals and value pricing

• Energy drinks, boba-style textures, new drink platforms

• Store efficiency budget raised from $3B to $4B, mostly new equipment

But comps went negative in the March 2024 quarter and kept sliding.

Niccol treated it as an operations problem. Make the store work better via:

• More labor hours and higher staffing floors

• A shorter menu

• A four-minute service target

• Ceramic mugs and a name written on the cup

That's where the operating profit decline was invested. And the numbers got worse as comps kept falling for roughly another year on his watch, six straight negative quarters in all, before Q2 FY2026 landed at +6.2% with transactions leading.

The reason I find this so compelling is that Narasimhan's read was the reasonable one that appears to be rational, fast, and measurable. Nicol’s approach is harder to quantify and takes longer to see the results.

A personal data point. I take a lot of meetings in coffee shops. For years, Starbucks was where I took all of them. Now I can't remember the last one. The stores got dirty. The seating turned into a holding pen for mobile orders. The whole place started to feel like a warehouse with an espresso machine in it.

The different approaches of these two CEOs make this one of my favorite reports. I feel like a lot of companies choose to spend more money on sales and marketing when there are likely more impactful, yet harder-to-quantify, opportunities for improving their core product or business.

With that, I’d love to see you join our upcoming Nerd Out workshop. And I’ll see you in a couple of days with another report.

Nick

PS. Starbucks needed a $1B restructuring and two years of ugly earnings to fix their throughput problem. Whatever's clogging your week is smaller than that and a lot more fixable. Bring it to the workshop and we'll build the first version of the fix together in 90 minutes. Registration closes at Noon CT tomorrow. NerdOut.com

TL;DR

• Starbucks runs roughly 41,000 coffeehouses globally, about half company-operated and half licensed, with 73% of revenue from the US.

• The business is ~61% beverages, and the economics live and die on how many drinks a store pushes through per hour.

• Brian Niccol's "Back to Starbucks" plan traded margin for throughput: operating margin fell from 15.0% to 7.6% while comps turned from negative to +6.2% in Q2 FY2026.

• CapEx is down 42% from the FY2024 peak, and in March 2026 they sold 60% of Starbucks China to Boyu Capital for about $4B while keeping 40% plus brand licensing income.

• Working capital runs negative $2.8B because roughly $1.8B of gift card and app balances arrive before anyone drinks anything. The customer funds the inventory.

• The stock returned +13.1% over the past year, ahead of McDonald's at -7.9%, but is still down 8.0% over five years. So far the market has paid for the recovery and stopped there.

The 30,000-Foot View

Starbucks makes money the way every restaurant does, one transaction at a time, but with two twists. First, the ticket is small and the frequency is enormous, so throughput is the entire game. A store serving 500 drinks a day at four minutes of wait is a completely different business than the same store serving 400 at twelve, on identical rent and equipment.

Second, they've built a payments and loyalty layer on top of the coffee. Rewards hit a record 35.6 million 90-day active US members in Q2 FY2026. Customers preload money onto the app, which hands Starbucks an interest-free balance and a read on demand before it walks in the door. That's the part most operators miss here.

FY2025 is what a self-inflicted rebuild looks like. Revenue grew. Operating profit got cut nearly in half. Then in September 2025 they took a $1B restructuring charge, closed 627 stores, and cut about 900 nonretail roles on top of 1,100 corporate jobs earlier that year.

Revenue mix (FY2025):

• Beverages: ~61%

• Food: ~19%

• Packaged coffee, licensing, merchandise, and other: ~20%

Geographic mix (FY2025):

• United States: ~73%

• International excluding China: ~19%

• China: ~8%

Key Stats

• Market cap: $117.7B

• TTM revenue: $38.5B

• TTM gross margin: 21.6%

• 1Y total return: +13.1%

• Employees: ~381,000 (before the China joint venture closed)

• Industry: Restaurants, specialty coffee

Company History

• 1971: First store opens in Seattle's Pike Place Market selling roasted beans and brewing equipment. Nobody was pouring espresso yet.

• 1987: Howard Schultz buys the company for $3.8M and starts converting it into an Italian-style espresso bar chain.

• 1992: IPO with 165 stores and a market cap around $270M.

• 1996: First store outside North America opens in Tokyo.

• 2008: Schultz returns as CEO mid-crisis, shuts every US store for three hours of espresso retraining, and announces 600 closures.

• 2018: Signs a $7.15B Global Coffee Alliance with Nestlé, handing packaged-coffee distribution to a partner while keeping the brand.

• 2021: A Buffalo store votes to unionize, the first of hundreds.

• 2024 (September): Brian Niccol arrives from Chipotle and launches "Back to Starbucks."

• 2025 (September): $1B restructuring, 627 stores closed, ~900 nonretail roles cut.

• 2026 (March): Boyu Capital acquires 60% of Starbucks China for about $4B. 7,991 company-operated Chinese stores convert to a licensed model, with Starbucks keeping 40% and the brand licensing stream.

Show Me the Money

Standout financial features:

• Operating margin compressed from 16.3% in FY2023 to 7.6% trailing, almost entirely from store operating expense. Every dollar of that got signed off on upstairs.

• Revenue grew the whole time, $36.0B to $38.5B, which is the tell that the demand side was never broken.

• CapEx is down 42% from the FY2024 peak of $2.8B to $1.6B trailing. They stopped building and started fixing.

• Working capital sat at negative $2.8B in FY2025, with $1.8B of that funded by stored value balances customers had already loaded.

• FY2025 dividends of $2.8B exceeded free cash flow of $2.4B, with net debt at $22.9B against negative book equity. That gap is worth watching.

Financial Data

TTM covers the four quarters through March 29, 2026. FY2025 includes roughly $1B of restructuring charges.

Stock Performance

Dividend-adjusted total return through July 24, 2026.

The one-year number looks like a turnaround being rewarded, and it is. Five years still sits negative, though. Anyone who bought in mid-2021 has ridden a full cycle of overexpansion, margin compression, a CEO change, and a rebuild just to end up behind where they started.

The N.O.O.B. Nine — Competitive Powers

The Nerd Out on Business Nine is made up of Hamliton Helmer's famous "7 Powers" of competitive advantage (Scale Economies, Network Economies, Counter-Positioning, Switching Costs, Branding, Cornered Resource, and Process Power) combined with two of my own (Data Flywheel and Distribution Advantage).

Average Score: 3.3/5 - A durable brand-and-distribution moat sitting on top of an operating system that turned out to be far more fragile than the brand suggested.

Memorable Marketing

Starbucks markets the store experience first and the product second. The FY2026 push is unusually literal about it. The ads put baristas and coffeehouses on screen and let the coffee stay in the background.

Notable campaigns and tactics:

• Handwritten cups (2025): Names and notes back on the cups, "for here" drinks in ceramic. Cheap, on-brand, and it made the operational change visible to customers.

• "Hello Again" (Super Bowl 2026): Barista routines set to AC/DC's "Thunderstruck." The whole spot is a reintroduction built on the rhythm of the work.

• "The Magic of Coffee" (2026): Directed by David Gelb of Chef's Table, leaning entirely into craft and the espresso itself.

• Rewards relaunch (March 2026): New Green, Gold, and Reserve tiers with perks like Free Mod Mondays, aimed squarely at visit frequency.

• "The Devil Wears Prada 2" collection (2026): Character-themed drinks tied to a film release, the modern version of the Pumpkin Spice Latte trick.

Tactical takeaways:

1. When you fix an operational problem, make the fix visible. The handwritten cup is a $0 signal that something changed.

2. Build loyalty around frequency and perks. Discount-driven programs quietly turn into a permanent price cut.

3. If your business is a place, market the place and the people in it.

4. Tie limited-time products to cultural moments you don't have to manufacture yourself.

AI Uses & Opportunities

Current exposure:

• Green Dot Assist, a generative AI assistant built on Microsoft Azure OpenAI, went from a 35-store pilot in mid-2025 to thousands of North American stores in early 2026. Baristas ask by voice or text and get recipe cards, ingredient swaps, and equipment troubleshooting without digging through a manual.

• Deep Brew, the older machine-learning stack, already drives demand forecasting, inventory, and labor scheduling.

• An internal "GROW Report" surfaces store-level performance gaps fast. Boring, and probably the most valuable of the three.

Future opportunities:

• Predictive maintenance on espresso machines, where one dead brewer wrecks a store's day.

• Mobile-order sequencing that stages drinks against predicted arrival time so nothing sits on the counter going cold.

• Churn and frequency modeling against a 35.6M-member Rewards base.

• Voice ordering at the drive-thru, a playbook Niccol already ran at Chipotle and Taco Bell.

Bumps in the Road

• The turnaround is exactly one quarter of clean top-and-bottom-line growth old.

• Labor is unresolved. A 131-day strike ended when bargaining resumed. There's still no contract, and staffing levels are the disputed issue.

• Net debt of $22.9B with negative book equity leaves little room for another expensive mistake, and the dividend already outruns free cash flow. That math works while CapEx is suppressed and gets harder once they start building again.

• China is now a 40% stake and a royalty stream. Better risk profile, less upside, and a lot less operational visibility.

• The stored value float is a liability that grows with customer trust and shrinks the moment people stop reloading. It has been rock steady, and it's worth watching if the turnaround stalls.

Your Swipe File

• Falling sales with steady brand awareness can points at operations needing improvement. Operational improvements can be less clear and more risky than things like marketing spend, but don't discount the power in making your core business better.

• Throughput is a margin line. Starbucks bought speed with labor hours, ate a 7-point hit to operating margin doing it, and got transaction growth back. Try to quantify what service time means to your business, if that's part of your business.

• Get paid before you deliver and your customers finance your inventory. Negative working capital pays you, literally.