Stand at the head of the hallway in a new suburban house at six in the morning and look down its length. The sun has not yet reached the front of the house, so the corridor runs gray from the master suite to the laundry room, past three closed bedroom doors, a linen closet nobody opens, and forty-two feet of wall whose only job is to move you between rooms you actually use. You have stood in this hallway a thousand times and never once been shown what it costs.
Every floor plan has this corridor, and every buyer pays for it by the square foot, at the same price as the kitchen, the same price as the primary suite, priced as if a hallway were a room. It is not a room; it is the tax you pay on the rooms. Nobody itemizes it.
The Arithmetic Nobody Prints
Methodology first, because the numbers deserve it: the National Association of Home Builders' 2024 construction cost survey puts the average new single-family home at $428,215 to build, about $162 per square foot, on an average home of 2,647 square feet, and that figure covers construction only, no land, making it the closest thing American homebuilding has to a national unit price.
Now the share, and here is the astonishment: no published US single-family benchmark exists for how much of a house is circulation, meaning hallways, stairs, foyers, entries, the square footage devoted to moving rather than living, and the design literature offers only scattered anchors. One architecture reference puts circulation at 25 to 40 percent of total floor area in most buildings, but that figure is commercial-weighted, swollen by lobbies, elevators, and egress stairs, and it would be absurd applied to a house. A housing-efficiency study out of Iraq measured 17 percent of ground-floor built area consumed by stairs and internal corridors alone in its sample of houses. Residential designers generally treat 8 to 10 percent as efficient and 12 to 18 percent as sprawling, figures I could not find published anywhere as a formal benchmark, which is itself part of the story.
Take 10 percent as the working middle, and on the NAHB's 2,647-square-foot average home that is 265 square feet of corridor, which at $162 a foot costs $42,900 to build. Stretch the range to 8 percent for a tight plan and 15 percent for a hallway-heavy one, and the hallway tax runs from $34,300 to $64,300. Those bands, under 8 percent excellent, 8 to 12 percent typical, over 15 percent hallway-heavy, are my own synthesis of the literature, not an industry standard. I am labeling them as such because nobody else will, and because a number with a stated uncertainty is worth more than a number pretending to have none.
For perspective, $42,900 is more than half the entire framing package on the NAHB's cost breakdown, which runs 16.6 percent of construction cost, so the corridors you walk through cost you over half of what the structure holding the house up costs.
The number moves with your market, because the corridor follows local labor rates the way everything else in a house does: NAHB's regional data shows custom-build square-foot costs running from about $129 in the East South Central states to over $190 in New England, with spec-built homes at $153 nationally. In a New England custom build, that same 265 square feet of hallway costs north of $50,000, while in Alabama it costs $34,000.
The Machine That Learned to Read a Floor Plan
Here is the part that should embarrass the industry. Machines can now see the hallway tax, and they can see it with an architect's eyes.
Training data arrived first, in 2019, when the RPLAN dataset put more than 80,000 human-designed residential floor plans into the hands of researchers, who promptly taught deep networks to generate new plans from layout graphs. Last year, ResPlan contributed 17,000 vector floor plans at true metric scale, with typed room-connectivity graphs recording exactly which rooms connect through doors, which merely touch, and which open to the outside. A connectivity graph is, among other things, a circulation map: it knows every path you can walk in a house.
Then came the judgment, at ACL 2025: the FloorPlan-LLaMa project built ArchiMetricsNet: 120,000 floor plans scored by professionals on three dimensions, functionality, circulation, and overall quality. Circulation, as a formal, machine-learned evaluation dimension, is younger than two years old. Architects have critiqued flow since Vitruvius, but no training dataset had ever asked a model to learn what good movement feels like in a house. Now one has.
Commercial money moved faster: Higharc raised a $90 million Series C in July for what it calls spatial AI: models trained on thousands of residential drawings that parse any floor plan image into a structured 3D model, identifying room types from features, an island means kitchen, and linking the result to a cost estimate. Signature Homes in Alabama fed it a plan and went from concept to permit-ready construction documents in about two weeks, work that traditionally takes six months, and a system that can identify every room from a sketch could trivially compute what share of a plan is hallway. It simply does not market that number, because builders are its customers and builders do not advertise the hallway tax.
Elsewhere, TestFit's generative design engine, launched in 2024, tests "every possible configuration" of a site, optimizing for efficiency metrics like floor-area ratio and yield on cost. Academic work has formalized quantitative methods for scoring circular movement through floor plans. So the capability is everywhere, and the scorecard is nowhere.
The Score That Does Not Exist Yet
Say this plainly, because the trade press will not: there is no consumer product, today, that scores a single-family floor plan's circulation efficiency for a buyer. Not from the AI startups. Not from the plan marketplaces. Not from your builder. Datasets exist, connectivity graphs exist, professional evaluation rubrics exist, and the buyer deciding between Plan A and Plan B gets granite countertop samples instead of the one number that would tell them which plan wastes less of their money on walking.
This is the gap the industry has every incentive to leave open. A circulation score would be awkward for the builder whose "luxury" model carries a 60-foot gallery hall. It would be awkward for the plan marketplace whose best-selling design eats 14 percent of its square footage on a two-story foyer. Metrics threaten margins, and so the metric does not exist.
But the raw material sits in your builder's CAD files, and computing it takes minutes, so ask for it. Ask anyway. Ask for the circulation ratio on every plan you are comparing: hallways, stairs, foyers, and entries, divided by finished square footage. A builder who has the number and shares it is a builder worth working with. A builder who cannot produce it, or will not, has told you something too.
An honest technology note is owed here, in the spirit of not overselling the machine. Everything in the previous section describes B2B and research systems; none of them is a product you can point at a plan PDF today and get a circulation score. My advice here is deliberately low-tech, because the high-tech version is still a pitch deck. When someone sells you an AI plan analyzer for your house, ask what dataset it trained on and whether a licensed architect validated its scores. If the answer is RPLAN and no one, you are buying a model trained on Asian apartment layouts to judge your suburban colonial.
How to Measure Your Own Hallway Tax
You do not need the AI; you need a tape measure, the floor plan, and ten minutes. That is the whole toolkit.
Measure every hallway, every stair run, every foyer, and every entry vestibule. Add them up. Divide by the plan's finished square footage. That is your circulation ratio. Compare it against the bands above. Then multiply the hallway square footage by your local construction cost per square foot, $162 nationally, adjusted for your region, and you have your hallway tax in dollars. Write it on the plan in red ink; it changes how the plan looks.
While you are at it, learn the four signatures of a hallway-heavy plan. First, the double-loaded corridor: bedrooms on both sides of a long hall, the cheapest way to connect rooms and the most expensive in square footage per bedroom. Second, the gallery hall: a corridor widened to eight or ten feet and rebranded as "art space," which is a hallway with better lighting. Third, the two-story foyer, which can easily swallow 150 square feet of air and call it grandeur. Fourth, the formal living and dining rooms you walk through to reach the rooms you use, circulation disguised as architecture.
None of these is automatically wrong, and that is the point of measuring rather than moralizing: a gallery hall with your actual art on the walls is a room that happens to be long. A 60-foot corridor to three bedrooms nobody visits is $29,200 of walking, and the number tells you which one you are looking at.
If you are choosing between plans, make the builder do the arithmetic for all of them, on the same basis, in writing, which takes five minutes in CAD. If two plans differ by three percentage points of circulation on a 2,647-square-foot house, that is 79 square feet, or $12,800 at national average cost, and you would negotiate harder than that over the appliance package, which at least keeps your food cold.
The Case Against
Against this thesis, the strongest argument deserves a full hearing, because it is mostly correct: hallways do work. They separate the noise of the living room from the quiet of the bedrooms, which anyone who has tried to sleep through a teenager's video games will appreciate. They give doors somewhere to swing without hitting furniture. They provide, in many plans, the only uninterrupted wall space for art and bookshelves. And the zero-hallway extreme is the open-plan great room, whose romance is already cooling as buyers rediscover that cooking smells, television noise, and homework all share the same air, and that a house with no doors is a house with no privacy.
There is a deeper objection, and it is the one I hold most seriously as someone who writes about design. Optimizing circulation to zero is exactly the kind of flattening that makes every AI-generated plan feel the same: the efficient plan, the rational plan, the plan with nothing in it that a model could not justify. Architecture has always spent square footage on things that do not pencil out. Consider the enfilade, that Baroque sequence of doorways aligned through a palace's rooms: pure circulation theater, and magnificent. A hallway is not waste by default; an unexamined one is. Know the difference.
I am not fully persuaded by the counterargument, for one reason: nobody is examining it. Buyers face a choice, but not between a thoughtful corridor and a ruthless optimization; it is between plans whose circulation was designed and plans whose circulation was leftover, and the buyer cannot tell which is which because the number is never printed. Measure the hallway, price the hallway, and then decide, with open eyes, whether yours is an enfilade or an accident, which is all this article asks.
Limitations
Several honest caveats, starting with the bands themselves: they are my synthesis of scattered literature, not measured data, no published US single-family circulation benchmark exists, and the 8 to 15 percent range could shift if one were ever compiled, while the $162 per square foot figure is a national average that moves from roughly $129 to over $190 by region and build type, so compute your hallway tax with local numbers. Hallways are cheaper to finish than kitchens and baths, so pricing corridor space at average cost per square foot overstates the true marginal savings of trimming it; the real number is lower, though still large. RPLAN's 80,000 plans are predominantly Asian residential layouts, and layout priors learned from them may not transfer to American suburban plans. Finally, no consumer circulation-scoring product was found at publication time; the "ask your builder" advice assumes a cooperative builder, and some will not cooperate.