opening a cafe — the real playbook

how much does it cost to open a coffee shop?

opening a coffee shop costs roughly $35,000 to $80,000 in the UK, $50,000 to $250,000 in major US cities, ฿800,000 to ฿2,500,000 in bangkok, and AU$70,000 to AU$200,000 in melbourne or sydney. the floor catches you when you take over a cafe that's already fitted. the ceiling hits when you build from a shell on a high street and want the espresso machine you actually want.

the four line items that dominate every budget are the same in every city: fit-out, equipment, deposit and three months of rent in cash, and working capital for the first ninety days. underestimate any of these and you'll be raising money mid-build, which is the single most common reason cafes never open after signing the lease.

the cheapest path to opening is the takeover. you find a tired cafe whose owner wants out, you buy the lease + fit-out + key equipment for one lump, and you re-skin the brand. takeovers in london's zone 3 trade hands for £25k-40k all-in. the trade-off is you inherit the lease terms, the kit you'd not have chosen, and whatever reputation the previous owner built or destroyed.

the most expensive path is the flagship shell: a unit nobody's drunk coffee in yet, a high street where the rent reflects footfall, and a fit-out that has to compete with the established names down the street. flagship shells in nyc's soho or williamsburg start at $150k and climb. you earn it back faster if you nail it. you lose it faster if you don't.

the 90-day pre-launch sequence

the day you sign the lease, you start a clock. ninety days from that morning is when you should be pulling shots for paying customers. anything slower bleeds rent and morale. anything faster usually means something important got skipped. this is the order that works.

the equipment that earns its space

your espresso machine is the single biggest decision you'll make and the easiest to over-think. the truth is, every machine on the working specialty list pulls a beautiful shot. what differs is workflow ergonomics, temperature stability under volume, parts availability, and how much your head barista wants to use it on a tuesday rush.

our shortlist for a new specialty cafe at 100-300 covers/day: la marzocco linea pb (the safe modern default, every barista has trained on one), victoria arduino black eagle (volumetric, paddle-driven, fast), slayer espresso (pre-infusion control, specialty darling), modbar AV (under-counter, opens sight-lines, polarising). for under-200 covers a nuova simonelli aurelia or a refurbished la marzocco linea classic saves $5-8k and tastes within 5% on most blends.

your grinder matters more than your espresso machine. an EK43 grinds filter and signature drinks brilliantly. a mythos one or mahlkönig E80 GBW handles your espresso. you want two grinders minimum — one for your house espresso, one for the guest single-origin or decaf. browse our full equipment directoryfor every machine and grinder we've indexed, with manufacturer specs and price ranges.

where to source your beans

you have three options. you can buy from a local roaster whose name your customers already know. you can buy from a roaster who's smaller and hungrier and will make your blend with you. or you can roast yourself, which means stop reading the cafe playbook and read the roastery playbook.

option one is the safest. the established roaster brings brand equity, training, and consistent supply. you pay a small premium and accept they probably supply twenty cafes within walking distance of yours. in london that's square mile, monmouth, workshop, ozone. in melbourne it's seven seeds, market lane, st ali. in tokyo it's onibus or fuglen.

option two is the path most independent cafes take after their first year. you find a roaster two years younger than you who roasts beautifully and isn't yet on every menu. they'll cup with you, develop a signature espresso, and grow with your business. start by browsing roasters who explicitly publish their wholesale terms — minimum order, lead time, certifications — in our wholesale directory.

hiring and training your baristas

your bar is only as good as the slowest person on it during a friday rush. you hire for temperament first, technique second. a coachable junior with three months of experience beats a rude senior with three years every time. the seniors who matter most are calm, technical, and present in a way the room feels. they cost £14-18/hr in london, $22-30/hr in nyc, ฿180-260/hr in bangkok, AU$28-36/hr in melbourne plus penalty rates.

training is where most cafes underinvest. SCA-licensed barista skills foundation + intermediate is the cleanest signal of seriousness for a hire. send your head barista on the SCA roasting professional + sensory skills foundation if you want them dialing in espresso the way the best in your city dial in. our coffee schools list covers every SCA-accredited training centre across the UK, US, europe, australia, and SEA, plus the cult-favourite alternatives like square mile training and counter culture training centers.

a rule that holds across cities: budget 40 hours of paid training before opening for every barista you hire. you're paying for the muscle memory you need on day one. owners who try to skip this open with a queue that frays, drinks that vary, and reviews that punish them for both.

the margin math you'll live by

a flat white sells for £3.50 in london, $5.50 in nyc, ฿95 in bangkok, AU$5.50 in melbourne. the ingredients cost 70-90p, $1-1.40, ฿20-25, or AU$1.20-1.50 depending on milk + bean prices. gross margin on a flat white is therefore 72-78%. retail bags of beans run 60-65% gross. food sits at 55-65% gross. averaged across a typical specialty cafe, blended gross margin is 65-70%.

your operating costs as a percentage of revenue come in roughly as: rent 12-18%, labour 28-35%, cogs 30-35%, everything else 8-12%. the numbers that move the most are rent and labour. cafes that signed bad leases in 2024 are the ones closing now. the cafes that survived 2020-2023 had landlord conversations, period.

breakeven for a 30-seat specialty cafe at london or nyc prices is roughly 140-180 covers per dayat an average ticket of £4.20 / $6.50. melbourne and bangkok are tighter on per-cover spend but lower on labour, so the breakeven cover count is similar at 130-170/day. you should be at breakeven by month 4-6 if your location and concept are right. cafes that haven't crossed it by month 9 usually never will.

the five mistakes that close cafes in year one

1. 1

   signing a lease without negotiating a rent-free period

   every landlord will give you a 2-4 week rent-free build-out window. most owners don't ask, sign the lease at full rent from day one, and bleed cash through fit-out. a four-week rent-free window on a london zone 2 unit is £2,400-£4,800 you keep. ask, every time.

2. 2

   buying a machine you can't maintain

   the slayer is gorgeous and almost impossible to service in johannesburg or jakarta. a la marzocco linea pb has parts within reach of any technician on three continents. your choice should weight parts availability above paddle aesthetics, especially outside the major cities.

3. 3

   under-staffing the first three weekends

   the cafes that fail loudest fail in their first three saturdays. wait times of 12+ minutes turn early curiosity into negative reviews. open with one more barista than you think you need. you can let people go in week 5. you can't un-write a 2-star review.

4. 4

   skipping the pos integration during fit-out

   shopify pos + square pos + lightspeed all take an afternoon to set up correctly and a week of band-aids if you skip it. card readers must talk to the till. the till must talk to your roaster's order portal. set this up before the espresso machine is installed and you save a fortnight.

5. 5

   ignoring the first 50 google reviews

   your first 50 reviews on google decide the next 5,000 walk-ins. you reply to every one — positive and negative — within 24 hours, by name, with specifics. cafes that hit 4.6+ on the first 50 build a flywheel. cafes that ignore the first 20 reviews and average 3.9 spend the next year clawing back up.

founders who got it right

the cafes worth studying aren't the ones whose openings went viral. they're the ones still open eight years later, training the baristas who go on to open theirs. some examples worth a visit before you build yours:

• · kiss the hippo, london — started in richmond in 2018 with one cafe and a vintage probat, now operates seven sites and roasts for a hundred cafes. profile
• · seven seeds, melbourne — opened in carlton in 2007 as one of melbourne's first specialty cafes, now a benchmark for melbourne coffee culture worldwide
• · nothing before coffee, bangalore — opened in 2018 with one site, now operates a national chain of 30+ outlets and roasts in-house. profile
• · fábrica coffee roasters, lisbon — opened in 2014 in a 30-square-metre alfama unit, expanded to two locations within five years, supplies cafes across portugal
• · nüde cafe, canggu — opened in berawa in 2022 with a 12-seat patio, now the canggu specialty benchmark

study the city you're opening in — browse our curated city guidesto find the cafes earning the queues, then visit them. order what they offer at the time they offer it. notice the workflow, the staff calls, the music, the smell. you can't copy them. you can learn what good looks like.

frequently asked questions