DrinkX
ENRU
Request
October 9, 20266 min readRetail

Barista or Machine: the Economics for a Chain of 50+ Points

15-25% margin versus 50-55%, the payroll of fifty sites, and the costs a single point doesn't have. Plus three cases where a person behind the counter still wins.

For a single point, the "barista versus machine" argument is settled by taste and investment. For a chain of fifty sites it is settled by payroll arithmetic, and the result is almost always the same. It is more interesting to work out exactly where the line runs and in which cases a person behind the counter still wins.

What changed in 2026

The economics of a coffee shop shifted on several fronts at once. VAT rose to 22%, and businesses on the simplified tax system with income above 20 million roubles a year now pay VAT. Coffee beans became 25-30% more expensive, and the average ticket in coffee shops grew 14% in January-October 2025. Meanwhile the number of purchases in coffee shops fell 7% in 2025, and sales of ready-made coffee in January-March 2026 dropped 4% year on year, according to Platforma OFD.

The net profit of a classic coffee shop is around 20-25% of revenue in a good month, and in current conditions often lower. For a self-service point it is about 50-55%. The difference is not magic: the second model has no salaries, social contributions or full-scale premises.

Comparison by line item

ItemPoint with a baristaSelf-service machineStart-up investment per pointfrom 2.5-3 million ₽ including renovation and furniturefrom 100 to 600 thousand ₽, professional models up to 1.5 million ₽Floor areafrom 10-15 sq. mfrom 0.5 sq. m, a power socket and waterStaff per point2-3 people in shiftsnone, only servicingOpening hoursper shift scheduleround the clockMargin15-25%50-55%Payback periodfrom 1.5 to 3 years10-12 months with a good location

How to calculate for a chain, not for a point

On a single point the payroll difference looks like "minus 150-200 thousand a month". On fifty it is already 7.5-10 million roubles every month, and a completely different conversation is built around that sum.

But a chain has cost items that a lone entrepreneur doesn't. Those are the ones to count.

Service department

Fifty machines without service turn into fifty problems. You need field engineers, a spare-parts warehouse and a maintenance schedule. A rule of thumb is one engineer per 25-40 machines in a city with decent logistics. Per point that is 5-10 thousand roubles a month, and that figure eats part of the payroll savings.

Replenishment logistics

Someone has to deliver beans, milk and cups. In a chain with baristas, the point's staff does it as part of their shift; in a chain of machines it takes separate routes. The denser the points, the cheaper the round. Machines scattered one by one across a region burn their margin on petrol.

Quality control

The most underrated item. As Alexander Kuzmin, head of RusHolts, told RBC Saint Petersburg (March 2026), when one large chain with 500 machines was audited, not a single machine fully matched the chain's own technology card, and no two had the same settings. With baristas, the same problem is solved by training; with machines, by central configuration and telemetry. Without that, a chain sells fifty different drinks under one logo.

Losses at the till

In chain audits, the gap between drinks prepared and drinks rung up reaches 10% and more. A machine with cashless payment closes this hole by design; at a counter with a person, only discipline and control can.

Where a person behind the counter still wins

A machine loses in three situations, and they should be acknowledged honestly.

High ticket and a complex menu. Signature drinks, alternative brewing methods, working with the guest: here a machine is no competitor. If your model rests on a 350 ₽ ticket, saving on a barista will destroy the product itself.

Peak throughput. Two baristas on two groups serve more drinks in the morning hour than one machine. At sites with sharp peaks, such as a railway station or a large business centre, it is sometimes right to install several machines rather than argue with this arithmetic.

Selling add-ons. A person offers a croissant; a machine doesn't. Some chains cover this with a display case nearby and an upsell on the screen, but conversion is still lower.

The hybrid model

In practice, chains increasingly stop choosing one of the two. The scheme looks like this: top locations with a high ticket keep a counter with a person, mid-range and night-time sites move to machines, and some points get both, an operator by day and a machine round the clock.

This solution requires one condition: a single drink standard across both formats. Otherwise the guest gets different coffee under one sign and draws conclusions that don't favour the brand.

Decision order

1.
Sort the chain's points by traffic and average ticket. It usually turns out that a small share of sites brings in most of the revenue.
2.
For each group, calculate the margin including payroll, contributions and the rent of that particular site.
3.
Add service, replenishment logistics and telemetry to the machine model. Without them the calculation comes out 10-15% too optimistic.
4.
Check the tax position. A chain of points that consolidates its revenue hits the new limits faster than it seems at the planning stage.
5.
Run the pilot not on the best sites but on average ones. The best will show a result that won't repeat later.

Conclusion

In a chain, the choice between barista and machine is decided not by taste but by the revenue of each site. A point that sells fewer than 80-100 cups a day often doesn't pay for a staff shift.

At the same time, machines bring service, logistics and quality control with them. Build them into the calculation from the start, otherwise the payroll savings will turn out 10-15% smaller than expected.

Frequently asked questions

At how many points do machines become more profitable than baristas

It isn't about the number of points but about the revenue of each. If a site sells fewer than 100-120 cups a day, keeping a barista shift is almost always more expensive than the margin it brings. A chain simply makes this difference visible.

How many engineers does a fleet of machines need

A practical rule of thumb is one per 25-40 machines in a dense urban layout. For scattered sites the figure falls to 15-20.

Can the same taste be kept across all the chain's machines

Yes, if settings are defined centrally and telemetry shows deviations. By hand this can't be sustained on a fleet of several dozen machines or more.

What to do about night hours

A machine works round the clock with no night-shift surcharges. At sites with night traffic, such as fuel stations, railway stations and hospitals, this is often the main argument for self-service.

More on equipment and service: self-service coffee machine for business, coffee machine fleet monitoring, DrinkX station.

How this works on the under-counter station — on the home page; short answers on installation, water and service are in the Short answers section.

For business

What a station would earn for you

We’ll run the payback on your numbers: traffic, cup price, rent.

Read next — retail
Barista or Machine: the Economics for a Chain of 50+ Points | DrinkX