Modern MBA

Case study — Food & beverage · 9 min read · 5 questions

Why donut shops out-earn Krispy Kreme

The thesis

Krispy Kreme did everything scale is supposed to reward. It automated a handmade product, turned the production line into the marketing, went public in 2000 and grew past a thousand stores. Each line could make 12,000 donuts an hour, so the constraint was never supply — it was finding enough mouths. So it went everywhere: supermarkets, gas stations, airports. And the more places you could buy a Krispy Kreme, the less anyone wanted one. The average franchise fell from $692,000 in 2008 to $377,000 in 2015, operating margin peaked at 9% and is now 1%, and the average store grosses about $119,000 a year.

Dunkin' won by treating the donut as a loss leader. Six times the stores, seven times the earnings per store, over 60% of US sales in beverages, and the word Donuts removed from its own name. The product that gives the category its name is the thing both chains are trying to sell around.

Which leaves the actual money at the bottom. A 650-square-foot mochi donut shop in Portland grosses $600,000 — five times the average Krispy Kreme — at a 40% operating margin. It gets there by engineering the product: baked rice-flour batter, no yeast, no proofing, no fryer, no hood vent, no gas, one employee running front and back, and a 20-minute batch cycle so it never overproduces. Two miles away a craftsman making genuinely better $5 donuts at 3am grosses $8,000 a month against $12,000 of rent. Product is not the moat. Product engineered around the cost structure is.

How do you think about this? 5 strategy questions this case raises and does not answer.
Read the comments, or add yours →

The statistics

$119KAverage annual revenue of a Krispy Kreme store
40%Operating margin at a single 650 sq ft independent
9% → 1%Krispy Kreme's operating margin, peak to today

By the numbers — swipe or use arrows

01Krispy Kreme earned the hypeAnnual revenue in millions of dollars, leading up to and just past the 2000 IPO.
Krispy Kreme earned the hype — Why donut shops out-earn Krispy Kreme$0M$250M$500M$750M$118M1996$133M1997$159M1998$181M1999$220M2000IPO$301M2001$394M2002$490M2003$649M2004Modern MBA
View data
Value
1996$118M
1997$133M
1998$159M
1999$181M
2000 IPO$220M
2001$301M
2002$394M
2003$490M
2004$649M
02Then it cannibalized its own franchiseesAverage annual revenue of a Krispy Kreme franchise, in thousands of dollars. Halved in seven years as the chain expanded.
Then it cannibalized its own franchisees — Why donut shops out-earn Krispy Kreme$0K$250K$500K$750K$692K2008$592K2009$486K2010$503K2011$538K2012$508K2013$477K2014$377K2015Modern MBA
View data
Value
2008$692K
2009$592K
2010$486K
2011$503K
2012$538K
2013$508K
2014$477K
2015$377K
03And the wholesale accounts walkedTotal active wholesale accounts, both types combined. Retailers tired of receiving stale leftover inventory to sell.
And the wholesale accounts walked — Why donut shops out-earn Krispy Kreme05,00010,00015,00013,160200812,291200910,919201010,786201110,636201210,09920139,73920149,9262015Modern MBA
View data
Value
200813,160
200912,291
201010,919
201110,786
201210,636
201310,099
20149,739
20159,926
04Nine percent was the ceilingAnnual operating margin before and after private equity. The 2016 takeover rebuilt the logistics. It did not fix demand.
Nine percent was the ceiling — Why donut shops out-earn Krispy Kreme0%2.5%5%7.5%10%1%20093%20104%20116%20129%20139%20149%20153%2021Re-listed2%20221%2023Modern MBA
View data
Value
20091%
20103%
20114%
20126%
20139%
20149%
20159%
2021 Re-listed3%
20222%
20231%
05Operating income, in the old era and the new oneKrispy Kreme operating income in millions of dollars. Revenue is more than three times higher today and the profit is a quarter of the peak.
Operating income, in the old era and the new one — Why donut shops out-earn Krispy Kreme$0M$20M$40M$60M$5M2009$12M2010$15M2011$26M2012$38M2013$46M2014$48M2015$41M2021$28M2022$13M2023Modern MBA
View data
Value
2009$5M
2010$12M
2011$15M
2012$26M
2013$38M
2014$46M
2015$48M
2021$41M
2022$28M
2023$13M
06Dunkin' has six times the footprintNumber of stores worldwide. Krispy Kreme figures are 2021 to 2023; Dunkin' figures are 2017 to 2019.
Dunkin' has six times the footprint — Why donut shops out-earn Krispy Kreme05,00010,00015,00012,5381,600Year 112,8711,709Year 213,1371,910Year 3DUNKIN'KRISPY KREMEModern MBA
View data
Dunkin'Krispy Kreme
Year 112,5381,600
Year 212,8711,709
Year 313,1371,910
07And seven times the earnings per storeAverage annual earnings of a store by chain, in thousands of dollars. Krispy Kreme figures are 2021 to 2023; Dunkin' figures are 2017 to 2019.
And seven times the earnings per store — Why donut shops out-earn Krispy Kreme$0K$250K$500K$750K$1,000K$733K$133KYear 1$743K$129KYear 2$766K$119KYear 3DUNKIN'KRISPY KREMEModern MBA
View data
Dunkin'Krispy Kreme
Year 1$733K$133K
Year 2$743K$129K
Year 3$766K$119K
08One 650 sq ft shop out-grosses the average Krispy Kreme five times overAverage gross earnings per location, in thousands of dollars. Mikiko is a single store in Portland with eight flavours on the menu.
One 650 sq ft shop out-grosses the average Krispy Kreme five times over — Why donut shops out-earn Krispy Kreme$0K$250K$500K$750K$1,000K$766KDunkin'$600KMikiko$119KKrispy KremeModern MBA
View data
Value
Dunkin'$766K
Mikiko$600K
Krispy Kreme$119K
09Where the margin actually comes fromOperating costs as a percentage of revenue. Mikiko pays above-market wages and still runs 20 points less labor, because one person runs front and back.
Where the margin actually comes from — Why donut shops out-earn Krispy Kreme0%10%20%30%40%33%26%Krispy KremeKrispy Kreme35%15%MikikoMikikoLABORFOODModern MBA
View data
LaborFood
Krispy Kreme33%26%
Mikiko35%15%
10The highest margin on the boardAnnual operating margins. A single independent beats every chain in the category and most of fast food.
The highest margin on the board — Why donut shops out-earn Krispy Kreme0%10%20%30%40%50%40%Mikiko33%Dunkin'24%Taco Bell24%Chipotle16%McDonald's16%Starbucks13%KFC9%Krispy KremeModern MBA
View data
Value
Mikiko40%
Dunkin'33%
Taco Bell24%
Chipotle24%
McDonald's16%
Starbucks16%
KFC13%
Krispy Kreme9%
11Three times the chains, for a dozenAverage price of a dozen donuts, 2024. Lola's prices at the top of the market on hand-made brioche, cream and glaze.
Three times the chains, for a dozen — Why donut shops out-earn Krispy Kreme$0$10$20$30$40$50$45Lola's$25Mikiko$14Dunkin'$14Krispy KremeModern MBA
View data
Value
Lola's$45
Mikiko$25
Dunkin'$14
Krispy Kreme$14
12Best product on the block, worst business on itLola's projected first-year gross earnings against the others, in thousands of dollars. Two months in, it grosses $8,000 a month.
Best product on the block, worst business on it — Why donut shops out-earn Krispy Kreme$0K$250K$500K$750K$1,000K$766KDunkin'$600KMikiko$119KKrispy Kreme$104KLola'sModern MBA
View data
Value
Dunkin'$766K
Mikiko$600K
Krispy Kreme$119K
Lola's$104K
13The rent alone is $4,000 more than the revenueLola's monthly cash flow at two months old. Rent covers both the mall storefront and the shared commissary twenty minutes away.
The rent alone is $4,000 more than the revenue — Why donut shops out-earn Krispy Kreme−$15,000−$10,000−$5,000$0$5,000$10,000$8,000Revenue−$1,600Product costs−$12,000RentModern MBA
View data
Value
Revenue$8,000
Product costs−$1,600
Rent−$12,000
01 / 13

Krispy Kreme revenue, operating income, operating margins, franchise average revenue and wholesale account counts from Krispy Kreme Doughnuts Inc. and Krispy Kreme Inc. filings for the years shown; Dunkin' store counts and average store earnings from Dunkin' Brands filings; store-level revenue, cost structure, margins, pricing and cash flow for Mikiko Mochi Donuts and Lola's Doughnuts disclosed on camera by their owners; comparative operating margins for Taco Bell, Chipotle, McDonald's, KFC and Starbucks from company filings; Lola's first-year figure is a projection from two months of trading

Key takeaways

01

Krispy Kreme rode a genuine product advantage into the IPO. Revenue went $118M, $133M, $159M, $181M, $220M, $301M, $394M, $490M, $649M from 1996 to 2004. People did not want a glazed donut. They wanted a Krispy Kreme.

02

The bottleneck was never production. Each line could make 780 to 12,000 donuts an hour, so every store had to hold equipment idle — which is why the chain chased volume into supermarkets, gas stations, convenience stores and airports rather than opening more of its own shops.

03

Ubiquity killed the novelty. The average franchise grossed $692,000 in 2008 and $377,000 by 2015 — cannibalized by its own expansion, halved in seven years.

04

Wholesale went the same way. Active accounts fell from 13,160 in 2008 to 9,926 in 2015 as retailers got tired of receiving stale leftover inventory to sell.

05

The margin tells the whole arc: 1%, 3%, 4%, 6%, 9%, 9%, 9% through 2015, then 3%, 2% and 1% after the private equity turnaround. Nine percent was the ceiling of a single-product company.

06

Operating income never cleared $48M in the old era and runs $41M, $28M, $13M now on $1.4B to $1.7B of revenue. The hub-and-spoke rebuild fixed the logistics. It did not fix demand.

07

Dunkin' solved it by making the donut a loss leader. It has 12,538 to 13,137 stores against Krispy Kreme's 1,600 to 1,910 — six times the footprint — with over 60% of US sales in beverages and the word Donuts stripped out of the name.

08

Per store the gap is worse than the count. Dunkin' averages $733K, $743K and $766K a store while Krispy Kreme averages $133K, $129K and $119K — roughly seven times the earnings per location.

09

A single 650-square-foot shop in Portland beats that. Mikiko grosses about $600,000 a year — five times the average Krispy Kreme — with over 90% of sales in donuts and a menu of eight flavours.

10

It wins on cost structure, not craft. Butter mochi and rice flour with no rising agent, piped into molds and baked in an electric oven — so no proofing cabinet, no fryer, no hood vent, no walk-in, no gas. The entire shop runs on a microwave, an oven and three mini-fridges.

11

The product stays good for days rather than hours, which changes the whole inventory problem. A batch takes 20 minutes from piping to finish, so the shop never overproduces and never runs out — while a conventional shop must guess in the morning and throw away what it guessed wrong.

12

Every step is measured. Batches of 80 — four trays of 20 molds. Start heating glaze when the oven has 10 minutes left. One pint of glaze tops 30 donuts; one 12-ounce bottle fills 30. Hold 10 of each flavour on weekday afternoons, 25 on weekday mornings, 40 on weekends. Customers are served in 30 seconds.

13

Even the sandwiches are an inventory play — the buns are made from unglazed day-old donuts, which is how a savoury item that raises check size also raises utilization.

14

The result is the highest margin on the board: 40% at Mikiko against 33% at Dunkin', 24% at Taco Bell, 24% at Chipotle, 16% at McDonald's, 13% at KFC and 9% at Krispy Kreme. Labor is 35% of revenue while paying above-market wages, food is 15%, rent is 5%.

15

Lola's is the control group. A craftsman making brioche, cream, glaze and frosting from scratch at 3am from organic flour and farm-fresh eggs, at $45 a dozen — three times Krispy Kreme and Dunkin' at $14, nearly double Mikiko at $25.

16

Two months in, it grosses $8,000 a month against $12,000 of rent and $1,600 of product costs. Fifty to sixty donuts a day, an $11 average order, first sale often after 1pm, mall hours from 10 to 9 after a 3am start, and production in a shared commissary twenty minutes away. Being good at your craft and being good at business are not the same thing.

Common questions

Are donut shops profitable?

The independents can be extraordinarily profitable, and the chains are not. A single 650-square-foot mochi donut shop in Portland runs a 40% operating margin — ahead of Dunkin' at 33%, Chipotle and Taco Bell at 24%, McDonald's at 16% and Krispy Kreme at 9% — with labor at 35% of revenue, food at 15% and rent at 5%. It grosses about $600,000, five times the average Krispy Kreme. But the margin comes from an engineered product and cost structure, not from craft: two miles of the same city has a shop selling genuinely superior $5 donuts that grosses $8,000 a month against $12,000 of rent.

What happened to Krispy Kreme?

It grew past the point where the product was special. Revenue went from $118 million in 1996 to $649 million by 2004 on the strength of an automated open kitchen and a distinct product people would drive to see. Because each production line could make up to 12,000 donuts an hour, the constraint was demand rather than supply, so the chain pushed into supermarkets, gas stations, convenience stores and airports. The novelty wore off as availability rose: average franchise revenue fell from $692,000 in 2008 to $377,000 in 2015, wholesale accounts fell from 13,160 to 9,926 as retailers received stale inventory, and operating margin peaked at 9%. Private equity took it over in 2016 and rebuilt it around hub-and-spoke production. Margin today is about 1%.

How does Dunkin' beat Krispy Kreme?

By not selling donuts. Dunkin' treats the donut as a building block for higher-margin drinks and sandwiches — over 60% of US sales are beverages, donuts are regularly given away free or packaged as promotional loss leaders, and the word Donuts has been removed from the company's name and logo. The scoreboard: roughly 13,000 stores against Krispy Kreme's 1,600 to 1,900, and $733,000 to $766,000 of average store earnings against $119,000 to $133,000. Both chains now run hub-and-spoke production, with centralized facilities baking and distributing to stores that receive, reheat and sell.

Why is mochi donut economics different from regular donuts?

Because it removes the fryer and the clock. Conventional donuts are yeast-leavened wheat dough that must proof, then fry, then be glazed fresh out of the oil — which needs hood vents, proofing cabinets, dry storage, walk-ins and gas, and produces a product that goes stale in hours. Mochi donuts are a rice-flour and butter mochi batter with no rising agent, piped into molds and baked in an electric oven. They peak in quality about six hours in and hold for days. That eliminates most of the build-out cost, lets one person run front and back, and cuts a batch cycle to 20 minutes — so inventory can be replenished against actual demand instead of guessed at in the morning.

Why do LA donut shops sell $1 donuts?

Because they are family businesses with no payroll. Los Angeles has around 1,600 independent donut shops, most of them no-frills neighborhood mom-and-pops run by Southeast Asian families who have been serving generic donuts in pink boxes for generations. There are no outside workers and no hourly wages, and they all use the same industrial commercial mixes, glazes and fillings — which makes production simple and cheap at the cost of taste. That price floor is what any modern shop selling a $3 to $6 donut has to justify itself against.

Why did Lola's struggle when the product was good?

Because everything other than the product was working against it. The shop sits in a mall, which brings foot traffic but contractually requires standard hours — so the owner starts baking at 3am and stays until the mall closes at 9pm. It has no ventilation or refrigeration, so production happens at a shared commissary twenty minutes away and the mostly finished donuts are driven in, which removes any flexibility on freshness through the day. Slow sales forced a cutback to flavours sharing one dough and one glaze, which left the display case sparse and undermined the impulse purchase the mall was supposed to provide. At $45 a dozen against $14 at the chains, the average order is $11 — two donuts — and the shop grosses $8,000 a month against $12,000 of rent.

Discussion

  1. Krispy Kreme did everything scale rewards: automated a handmade product, made the line the marketing, went public and passed a thousand stores. The more places you could buy one, the less anyone wanted one. When does availability destroy value?

  2. The average franchise fell from $692,000 in 2008 to $377,000 in 2015 and operating margin from 9% to 1%. At which point in that decline should the strategy have been reversed, and what would have stopped it?

  3. Each line could make 12,000 donuts an hour, so the constraint was never supply but demand. What happens to a company that solves the wrong constraint extremely well?

  4. Independents out-earn the chain on the same product. What are they doing that cannot be systematized?

  5. You are handed Krispy Kreme. Do you shrink deliberately, and what would the market do to you if you announced it?

Related case studies