Zoning gets confused with the floor plan all the time - I hear it on almost every first call with an owner. They are two different things. The floor plan answers one question: how many tables fit. Zoning answers another: how the room shapes a guest's behavior - where they sit, how long they stay, what they order, and whether they want to come back.

That is why two rooms of the same size, with the same number of covers and the same kitchen, can bring in different revenue. And the cause is not the chef and not the location - it is exactly how the tables are arranged and where the guest goes once they have walked in.

Let me start with an idea that runs against intuition: the spot a guest likes most and the spot that brings in more money are often two different spots. From there we will get into why that happens, where you can see it, and what to do about it.

Tight seating is the most expensive saving in the room

One of the most common zoning mistakes looks like a concern for revenue. I get the urge to fit in more tables, but in practice more tables does not equal more checks - and below I will show why it turns against the owner.

A guest reads the distance between tables not as centimeters but as personal safety. When tables sit right on top of each other, around 70% of guests would rather be reseated (Robson, Kimes, Becker, Evans, Cornell Hospitality Quarterly, 2011). And the buffer you need is bigger than the drawing suggests: in that same survey, one in three still felt cramped even at a gap of roughly 60 centimeters. The discomfort hits hardest on a date and among women - which is precisely the occasion a guest chooses a restaurant for, rather than delivery.

This is not a matter of "like it or don't." Tight seating damages the most valuable thing you have - the return visit. In a field experiment at a New York restaurant, guests at closely spaced tables were noticeably less satisfied, and their willingness to return came out lower than for those seated with a normal gap: 6.16 versus 6.51 on the survey scale (Robson & Kimes, Cornell, 2009). The gap looks small, but it is not one lost cover. It is the entire future stream from a person who never came back a second time - and who, on top of that, did not recommend the place, did not bring friends, family, or acquaintances. And a restaurant's core revenue comes precisely from returning guests and the people they bring.

I regularly walk into rooms where the whole evening atmosphere has been sacrificed for four extra tables. Those four tables add revenue in the moment and subtract it over the span of a year. What you should be counting is not "how many fit" but "how much a guest can tolerate and still want to come back."

The best table for the guest is not the best for the owner

Now the counterintuitive part that almost no one runs the numbers on. The tables a guest loves most - the snug booths and banquettes against the wall - often bring in less money per seat than the tables they walk past.

The mechanics are simple once you stop measuring by average check. In a cozy booth the guest is comfortable, in no hurry, and stays longer. At an open table in the aisle they eat faster and leave faster. Converted into revenue per minute of seat time, the open table is often more profitable than the cozy one - simply because it turns over faster (Kimes & Robson, Cornell, 2004). This is data from a single restaurant, and the authors honestly flag the finding as directional (it points the way rather than guaranteeing the size) - but in our experience it matches what we see on real projects.

The key here is the right unit of measurement. Not average check and not turnover on their own, but RevPASH: revenue per available seat hour (Kimes & Chase, Cornell, 1998). It joins both quantities - how much a guest spent and how long they held the table. A room where everyone lingers and spends a lot, and a room with fast turnover and a modest check, can produce the same RevPASH by different routes.

So the job of zoning is not to make every table cozy. That is a trap: a room built only out of "pleasant" seats underearns. The job is to assemble the right mix - some seats work for dwell time and average check (couples, groups, the long dinner), some work for turnover (solo diners, the quick lunch, the guest who is here for half an hour). You can only see this by counting revenue per seat, not per guest.

A real case: 53 tables out of 56 - all for four

It sounds abstract until you apply it to a specific room. The Chevys chain once took its main 230-seat room apart bolt by bolt (Cornell case, Kimes & Thompson, 2004). It turned out that 53 of the 56 tables were built for four, while more than half of the guests arrived as couples or on their own. The mismatch between the tables and the real party size held peak occupancy at just 50% - the room felt "full" and was half empty in money terms.

Without touching the walls, they changed the table mix: instead of an almost solid wall of four-tops, it became 39 tables for two, 35 for four, 2 for six. They relocated the service stations and added staff to the shift. The redo took 10 nights, and the restaurant never closed.

The result two months later: peak occupancy rose from 50% to 59%, RevPASH went from $5.85 to $6.32, the net revenue gain against the control restaurants came to +5.1%, and payback ran about 11 months. Same room, same kitchen, same square footage. All that changed was the table composition - that is, the zoning.

This is evidence-based design in its purest form: not "it feels cozier now," but a measurable gain from a decision that was calculated before anyone moved a single piece of furniture.

Why some tables have a waiting list and others sit empty

If there is a table in the room that no one chooses of their own free will, that is almost never chance and almost never bad lighting. A bad table even has a name. At the legendary New York Four Seasons, the far, cut-off end of the room was called "Siberia" - it was where you seated the people you did not need to put on display. And in the prestigious part of that same room the very term "business lunch" was born: the space was designed for a deal-making conversation, and it created a new kind of guest. One room, two poles - and both were made by the layout, not the menu.

Behind "Siberia" there is a mechanism, not snobbery. A guest picks a seat by a very old rule: see the room, but keep your own back covered. In studies of seat preference, "prospect" turns out to be a significant predictor of choice (Mumcu, Duzenli, Ozbilen, 2010); the "prospect plus refuge" principle itself was described back in 1975 by Jay Appleton as an evolutionary frame: out in the open a person feels exposed, against a wall with a clear view they feel calm. People choose seats with the most control over their personal space, including through a physical barrier - a wall, a partition, the back of a banquette (Robson, Journal of Environmental Psychology, 2008).

So "Siberia" is not just the table by the restroom. It is also the open table in the middle of the room, with no wall, window, or banquette behind it, sitting on the waiters' route. A "weak" table is not a defect but predictable behavior. On a project you do one of two things with it: either strengthen it - with a partition, greenery, a high back, a change in floor level, to give the back support and the eyes a view - or honestly hand it over to fast turnover. What you cannot do is drop a table there "just to have one" and expect it to perform like a seat against the wall.

Density, turnover, and the conflict of goals

Density comes with a subtlety that contradicts the first section only at first glance. Cramped is bad. But empty is bad too. Across a large body of real orders, a moderately high room density was associated with higher spending on food, especially among guests who came in alone (Tu et al., Appetite, 2023). A half-empty room reads to a guest as "something is off here," and they hold back on their own.

There is no contradiction - there is a difference between "lively" and "cramped." A lively room with normal gaps works for revenue: the guest feels life around them, relaxes, and orders more freely. A room where the tables are jammed together works against the return visit. The line between these two states is exactly what zoning is about: dense in the sense of life, not dense in personal space.

And the main conflict, one you have to spell out before designing rather than after opening: guest dwell time and fast turnover are different business models, and the zoning for them is opposite. A wine bar, a chef's-menu dinner, the "make an evening of it" restaurant earn on time at the table - they need comfort, barriers, soft furniture. A coffee shop, a bistro, a business lunch earn on turnover - they need hard seating, throughput, a fast cycle. The very same move raises revenue in one format and cuts it in another. So the first question is not "how many tables" but "what should the guest do in this zone: linger, or free up the seat."

A room is several different zones, not one

A good room is not an even grid of tables but a set of zones with different jobs, which the guest moves through like a route. Danny Meyer, one of the best-known restaurateurs in the world, put this well: he deliberately divides the room into "several small communities within a larger space" so that the guest feels anchored rather than lost in a big hall ("Setting the Table", 2006). And he adds the part that matters most for the money: the main reason a guest wants to come back is the feeling that they are seen and recognized here.

In practice there are usually several zones. The entrance and decompression zone - the first few meters, where the guest is met and led in. The anchor seats - against walls, in nooks, with a clear view and support at the back: this is where you seat the people who came to stay a while, and this is where the average check runs higher. The fast-turnover zone - open tables and "half-hour" seats, working on RevPASH through speed. And separately - the format for solo diners: the bar counter and the communal table.

Solo diners are often underrated, and for no good reason. According to OpenTable data, a guest who comes alone spends on average almost half again as much per head, and demand for seats at the bar counter and high tables is one of the fastest-growing in the room. From the same source: about three quarters of guests say the ability to choose their table or zone would improve the experience of their visit (OpenTable, 2024-2026 data). The bar counter is not "where you seat someone when there is no table free," but a full-fledged revenue zone with its own guest.

What else matters in the layout of the room

I named the entrance zone among the others, but it deserves its own conversation - almost everyone uses it wrong.

The first two or three meters barely register with a guest. For the first few steps into the room they still have not switched over: they are shedding their street mode, looking around, adjusting to the light and the sound, not taking in details. In retail this is called the decompression zone - everything displayed in those first meters passes right by their attention (Paco Underhill, "Why We Buy").

Two consequences follow. First: do not give the entrance over to your best tables - right at the threshold they function as pass-through seats, and the guest reaches them already "switched on." Second: whatever is meant to create the first impression - the key view, the display case, the open kitchen, the wine cabinet - is placed not at the threshold but a little deeper, where the guest has already begun to look. The entrance gets not a punchy accent but a soft way in.

This is part of a general rule: the impression of a room forms before the menu, and even before the guest has sat down. How scent shapes it - right at the threshold, in 20-30 seconds - is covered in restaurant scent marketing.

How to run the numbers on zoning in your own room

There is no universal "correct" zoning - there is the right zoning for the format and for the specific space, with its windows, columns, and entrance. The exact numbers come not from someone else's study but from measuring your own room.

On our projects we count RevPASH not as an average across the room but zone by zone: where revenue per seat is higher, where a zone sits idle, where the guest leaves earlier than they could have. We measure this on site - the real time at the table for each zone, not an eyeball estimate and not a lump total from the till. From there the table mix is reworked toward the goal - exactly as in the Chevys case, only for your room and your format.

We test this principle constantly on our own projects - we go back into finished spaces and watch how the zones behave in the wild. The center of the room almost always confirms the rule: a guest is pulled toward walls, windows, and partitions, and they walk around the open middle. A table in the middle of the room loses by default - the back has nothing to lean on, and that is an instinct, not a whim.

We learned to solve such a zone not with a barrier but the opposite - with attraction. Since the central table has no protection at the back, we give it what the wall seats do not have: a reason to sit right here. We turn the middle into the visual center of the room - an art object worth taking that table for and worth photographing. This works on two things at once: on the seating itself, and on the fact that guests photograph the spot and post it - the room gets a stream of pictures from the very point that usually sits empty.

That is what we did on the Izakaya Yosai project: a central table for 16 carved out of concrete in the shape of a wave, with an installation suspended above it made of branches gathered along the banks of the Om and the Irtysh. A spot that by every rule should have sat idle became one of the most in-demand zones - people take that table on purpose, for the view and the photograph. The weakest table by instinct we turned into the most noticeable one, without putting up a single partition.

Without measurement, zoning stays a matter of taste: "let's put a little sofa here and a counter there." With measurement it becomes a decision with a clear return - you can see which zone earns and which one is listed in the plan but not in the revenue.

A room does not have to be cozy at every point - it has to work. Some tables hold the guest and lift the check, some turn over fast, some cover the solo diners, and all these roles are needed at once. The real mistake is not "an unlucky table" but the absence of intent: when the arrangement is done by eye, no one knows which zone earns and which one just takes up floor space.

Zoning is not a pretty diagram on a plan. It is when every square meter of the room has a job - and you can check that job with a number. The simplest place to start: count what each zone of your room earns right now.