I've been walking the streets of Barcelona in the evenings for years, and I keep seeing the same thing. There are bars you want to leave fast - you have a drink and you go, nothing holds you. And there are the ones you want to stay in. Order another round. Talk a little longer. Spend more.

For a long time I thought it came down to the concept, the food, the staff. Then I started to notice: very often it comes down to the light.

Light is read before anything else - before the materials, before the drinks, before you've even worked out whether the seating is comfortable. And lighting is one of the few areas of design where you have not "impressions" but real research with measured effects. On why we lean on data rather than taste in the first place, see the article on restaurant design without data.

Why light steers guest behavior at all

Let me start with what isn't up for debate - the physiology of the pupil. Bright light contracts the pupil: a person becomes more focused, more rational, faster. Dim light expands it: attention softens, the guest is in no hurry, decisions come more easily. "Let's have one more glass" is partly physiology too, not just pleasant company.

That's why 2700-3000 kelvin isn't a fashion but almost a standard for the evening format: that warm light reads as "evening, you can relax." And it's also why big chains like McDonald's deliberately do the opposite - cooler and brighter. That isn't bad taste, it's an operational decision: a highly alert guest eats faster, leaves faster, turns the table over more often.

The trouble starts when an evening restaurant copies the bright, evenly lit room of a fast-food place without noticing that this is the opposite business model. Light should work for your economics, not someone else's.

The subtlety almost everyone misses: it isn't about darkness

"Make it darker" is half the truth, and the dangerous half. In the classic experiment by Baron, Rea and Daniels (Motivation and Emotion, 1992), what lifted people's mood wasn't just any dim light, but light that was warm and dim at the same time. Dim but cool light - and, conversely, warm but too bright - had the opposite effect: mood dropped.

In other words, what wins isn't "darkness" but the combination of two parameters: a warm spectrum plus moderate brightness. Get one of them wrong and the intended "cozy half-light" turns either into "gloomy and uninviting" or into "bright and cheap." Most unsuccessful rooms aren't "too bright" or "too dark" - they're the wrong combination of spectrum and brightness.

Light changes more than mood - it changes taste

Even less obvious: light affects how a guest perceives the food and the drink itself - and how much they're willing to pay for it. In a study by Oberfeld and Hecht (Journal of Sensory Studies, 2009), around 500 people tasted the very same wine under ambient light of different colors. Under red light, the same wine seemed almost one and a half times sweeter, earned a higher overall rating - and for the same bottle of Riesling people were willing to pay about a euro more than under green light. The wine didn't change by a single drop - only the light around the glass did.

The same principle works on the plate. A dish looks "expensive" not only because of the ingredients and plating, but because of how the light falls on it: a warm beam with the right shadow adds depth and appetite appeal, a flat wash strips them away. The chef puts in the work in the kitchen, and half the impression is decided by the lamp over the table.

The key technique: not "dim everything" but gather the light into pools

Here's the technique that separates an atmospheric room from a merely dark one - and it works differently from what you'd expect. The study by Wu, He, King and Mattila (International Journal of Hospitality Management, 2021), with its telling title "In darkness we seek light," showed that what draws people in most strongly isn't a uniformly dimmed room, but overall half-light with pools of accent light on the tables. It's exactly this contrast - a dark background and a bright table - that reads as "beautiful" and "I want to be here." And it works especially strongly for people who've come on a date or for a close conversation.

The authors explain it through a sense of privacy: a pool of light over the table creates a zone of "your own" within the shared room. The guest lands on a small lit island and subconsciously feels it is theirs.

And that's the answer to the question in the title. In the places you want to leave, the light is usually even and general - either bright and flat, or simply dark. In the places you want to stay, the light is gathered: the background is dimmed and the table sits in a warm pool. The guest doesn't analyze this. They just feel that this is their place.

Three things an operator feels but has never said out loud

First. Bad light never shows up in reviews. Never. A rude server - they'll write about it. A cold dish - they'll write about it. Wrong lighting creates something else: the guest can't tell what exactly is off, but feels "not quite my place." They leave four stars, write something neutral, and never come back. It's the quietest way to lose regulars - you'll never learn the reason. You change the menu, hire a new manager, and the problem was in the lamps.

Second. The materials - marble, oak, warm concrete - the operator chose in a showroom under bright daylight or 4000K overhead fixtures. In the room they work under the 2700K of evening light. That's essentially a different material: a warm spectrum with a side-angle source opens up texture and depth, a cool one drains the life out of it. That's why the sample from the catalogue and the same stone in the room feel different - and it's a problem not of the choice, but of the light.

Third. Most restaurants run all day on a single lighting mode: one group, one button, from lunch to close. But the guest lives by a circadian rhythm - it feels natural to them when evening light differs from daytime light, and late evening from early evening. Today, dimmers, timers and a couple of independent groups are enough for this (it's a settings job, not construction). It isn't complicated or expensive. But almost no one does it.

What this is worth in money - honestly

Now about money, because this is usually where people start inventing percentages. I won't hand you a figure that "light adds so much to the check" - in any rigorous form, it doesn't exist. And that isn't an excuse, it's an honest reading of the data: in some lab studies dim light noticeably lengthens the meal and raises spending, in others (Food Quality and Preference, 2019) there's no consistent effect on how much is eaten. The effect has a direction, but its size depends on the format, the guest and the room - there simply is no universal percentage, and anyone who names an exact figure of "light = +X%" made it up.

What definitely exists is a direction, confirmed by the research above: warm, gathered light relaxes, lengthens the evening and raises willingness to pay (like the +€1 for the same wine in Oberfeld); even, bright light speeds people up and pushes them out. We don't guess the size of that effect in your specific room - we measure it on site: we look at how guests actually behave across the zones and how long they sit at the table, and tie it to the revenue - and you can see how the picture shifts when you touch the light. That's the difference between "it feels cozier, I guess" and "guests stay longer in the evening, and it shows in the revenue."

At the same time, light is one of the cheapest interventions possible in a restaurant. Changing the lamp spectrum, adding accents over the tables, installing dimmers - that's days of work without tearing into the finishes. Few other things where such a small investment so noticeably affects how a room feels.

What to check in your own room right now

Look, here are the questions I usually start with when I look at someone else's space.

Spectrum. Take a bulb from any fixture and check the marking. 4000K and above in an evening format is already a reason to talk. Not a verdict, but a reason.

Lighting groups. One button for the whole room is the bare minimum of infrastructure. There's no point talking about scenarios until you have at least two independent circuits: general background and accents on the tables.

Pools on the tables. Look at whether each table has its own pool of light - or whether the room is flooded evenly. An even wash is the first sign that you're losing the very "your own place" effect that keeps a guest staying longer.

The room at seven in the evening. Not at three in the afternoon and not in photographs - exactly when the evening atmosphere is supposed to be working. How different does it feel from lunchtime? If it's almost the same - that's a signal.

Reviews over three months. Don't look for the word "light" - it won't be there. Look for "somehow not our place," "the atmosphere was a bit off," "nice, but no urge to come back." That's the trace of wrong light.

You can see how this looks on completed projects in our projects.

A room doesn't have to be dark or bright. It needs to be put together around what the guest is meant to feel. A warm spectrum, moderate brightness, pools of light on the tables and different scenarios for day and evening - four levers that cost little, yet decide whether the guest leaves after twenty minutes or stays for a second glass.

Light usually lives at the end of the budget, as an add-on. And yet it's the cheapest way a restaurant has to steer guest behavior. It's a shame to waste it.