The Unit Mix Is the Business Plan

A floor plate is a spreadsheet with walls. Every label you put on a rectangle, studio, one-bedroom, two-bedroom, gym, is a line of income or a line of cost, and the plan is where those lines fight for the same square feet.

When I was drawing test fits by hand, it took me too long to understand that the developer wasn't asking me for a layout. They were asking me to prove a business plan would fit. The mix is the business plan. The plate is where you find out if it's true.

One floor-area budget, many homes

Start with the simplest version. My unit-mix study treats net rentable area as a fixed budget, with circulation and cores already taken out. Studios are 450 SF at $1,250 a month. One-bedrooms are 650 SF at $1,550. Two-bedrooms are 950 SF at $2,100. Income assumes 95% collection and a 35% expense share.

Those are illustrative numbers, and the study says so. But even illustrative numbers make the trade visible. Push the mix toward studios and you get more homes and more rent per foot. You also get more kitchens, more bathrooms, more plumbing runs, more doors, more parking demand and more management work. Push toward two-bedrooms and the opposite happens.

There is no proven optimal mix. Anyone who tells you there is has a market in mind and isn't telling you which one. What there is: a way to test whether the requested proportions fit this site and this form, and whether the rent and cost assumptions behind them survive contact with a real plan.

Try it: drag the requested counts for each home type and the area budget, and see how many whole homes actually fit and what they earn.

Rent per foot lies if you let it

Rent per square foot is the most quoted number in multifamily, and it's the easiest to misuse.

A smaller unit usually commands more rent per foot and less rent in total. Which one matters depends on how much plate each unit eats. My rent-per-foot study used a 200 × 65 ft double-loaded plate with a 5 ft corridor, 30 ft unit depth and $200/GSF of hard cost. Unit width is just the target area divided by depth. Whole bays fit on each side, and the leftover at the ends earns nothing.

That leftover is the trap. Change the unit size by a few feet and the number of bays that fit on a 200 ft side changes in steps, not smoothly. A "better" rent per foot can lose a whole unit at the end of the bar. Compare total rent alongside rent per foot, or the ratio quietly replaces the business case.

The corridor is the room that pays no rent

Nobody signs a lease for a hallway. The corridor uses floor area, structure, finishes and lighting, and earns nothing directly. That doesn't make it waste. It's how people get home and how they get out. But it does make it the most important room to size on purpose.

The corridor study draws one floor twice, with the same homes in both: a double-loaded bar and a point block, a compact tower with homes wrapped around a central core. Each home is 24 ft wide and 30 ft deep, 720 ft². The core is two stairs, two elevators and a lobby in one home-sized bay. Rent is an assumed $36 per ft² a year.

Here's the bar:

• 10 homes, 6 ft corridor. The corridor runs 144 ft and takes 864 ft², 86 ft² per home. With the core, the floor gives up 1,584 ft² and is 82% rentable. That's $57,024 a year of rent that never gets earned.
• Same 10 homes, 9 ft corridor. Now the corridor is 1,296 ft² and the floor is 78.1% rentable. $72,576 a year unearned. Three extra feet of hallway cost $15,552 a year, on one floor.

Now the point block:

• 4 homes. No corridor at all. Every door opens on the core lobby. 80% rentable, against 71.4% for a bar holding the same 4 homes.
• 6 homes. Still no hall. 85.7% rentable against 76.9% for the bar, a gap of $20,736 a year on one floor.
• 8 homes. The faces of the core run out of door frontage and a 3.5 ft hall wraps the core. It's 427 ft², still less than the bar spends. 83.4% against 80%.
• 14 homes. The hall stretches 13 ft out from the core to reach every door and grows to 1,296 ft². The point block drops to 83.3%. The bar, at 84.3%, wins. A second core gets you 81.5%.

That limit of 8 isn't a law of nature. It falls out of one assumption: every home needs about 10 ft of hall or core frontage for its door. Change that assumption and the crossover moves. That's the point of modeling it instead of arguing about it.

Try it: drag the end of the bar to add homes, its corridor edge to widen the hall and the slider to change rent, and watch the point block win and then lose.

Two cautions, because the study states them and so should I. None of this establishes code-compliant corridor widths, travel distances or unit layouts. And a plan that saves area by making routes confusing has just traded one kind of efficiency for another. Net-to-gross is a comparison tool, not a grade.

A label changes the program, not the footprint

It's tempting to think of unit mix as labels on slots. Studio here, one-bedroom there, swap them in the spreadsheet and you're done.

My labeling study does exactly that on purpose: eight equal rectangular slots, with area targets of 500, 650, 850 and 1,100 ft² and rents of $1,400, $1,800, $2,300 and $2,800 a month. The schedule happily reports those rents even when a slot is too small for its label. Relabeling doesn't make a feasible apartment, and it doesn't make revenue.

A real relabel touches walls, plumbing, windows, parking and access. Some plans are more flexible than others, and that flexibility should be described precisely: which walls can move, which wet walls can't, what still has to change. "Flexible" without that list is a sales word.

Amenities are a concession, until they aren't

An amenity is area that usually doesn't earn rent directly but makes the rent-earning spaces easier to lease. A fitness room might be the cost of entry in one competitive set and a real differentiator in another.

My amenity study holds 15% of gross area for access, divides the rest into equal unit-area budgets and amenity, and counts whole units. Baseline rent is $1,800 per unit per month. The question it answers is blunt: what premium does the amenity have to earn to replace the scheduled rent of the units it displaced? That break-even leaves out the amenity's construction cost, operating cost, vacancy and lease-up, so the real bar is higher.

Two things make amenity math hard. Competitors copy, so this year's differentiator becomes next year's baseline. And faster lease-up can justify a space even when ongoing rent doesn't. Both are reasons to put the amenity case in the model, in numbers, before anyone falls in love with a rendering of the roof deck.

Simulate the residents, not just the rent

A pro forma is a low-resolution simulation of occupancy. Rent, vacancy and expenses summarize assumptions about people using the building. A test fit simulates something else: geometric yield. How many homes fit, and how the area gets used. I wanted to see what happens when you raise the resolution on both.

The study draws one floor of a 160 ft double-loaded bar: 30 ft deep units either side of a 6 ft corridor, and a 24 ft core with the elevator on one side. Studios are 14 ft wide, one-bedrooms 22 ft, two-bedrooms 32 ft.

• 13 one-bedrooms, core at the end. 8,580 SF rentable, 10 ft of wall left over. A flat 5% vacancy collects $281,580 a year.
• Same plan, unit by unit. Give each resident two assumed dislikes: sharing a wall with the core costs 6 points of occupancy, and every full 40 ft between the elevator and the front door costs 2. The same floor now collects $272,004, $9,576 less.
• Core moved to the middle. The test fit doesn't notice. Same 13 units, same $281,580 flat. Unit by unit it collects $275,196, because the longest walk falls from more than 120 ft to under 80. Only the finer model can tell the two plans apart.
• Core slid to 12 ft from the end. The run beside it is too short for a 22 ft unit, and only 12 fit. Flat income falls to $259,920. Losing one unit costs $21,660 a year, more than six times what moving the core gained.
• All studios. 19 fit with the core in the middle: $303,240 flat, $297,864 unit by unit. Put the core back at the end and a 20th fits: $319,200 flat and $309,792 unit by unit. Longer walks, unhappier residents, more money.

Try it: drag the core along the corridor, tap units to swap studio, 1BR and 2BR, and flip between the flat and unit-by-unit vacancy models.

Residents matter. In this study, though, geometry wins big and behavior wins small, and you need both models to know which one you're looking at. The dislikes are assumptions, not measured behavior. Leasing and turnover recorded unit by unit are what would test them.

The stacking plan is the business model in section

Stack the floors and the picture gets bigger. A stacking plan shows floors and spaces next to tenants, lease rates, expirations and options. My stacking study keeps it simple: fully rentable plates, stated teaching rents, 35% expenses and a 6% capitalization rate to turn income into value.

What the stack exposes is concentration and timing. Which floors are occupied, when leases roll, and what a change of use does to access and services. Moving a tenant in the stack isn't just a planning move. It changes the cash-flow profile and the leasing risk. The plan holds the spatial half of that conversation. Lease terms and market assumptions still live in the financial model, and the two need to be looking at the same version.

Developers write the requirements

All of this makes the developer look a lot like a product manager. They decide what the building offers, who it serves, and how the constraints fit together. The architect translates and tests those requirements, and the good ones push back.

That push-back only works with numbers. My study ran 572 unit mixes on one building. When the brief calls for an amenity, the architect can show what it costs in units. When the plan exposes a costly constraint, the developer can revisit the brief. Neither side gets to be right in the abstract.

Try it: hover the 572 mixes, toggle premium and standard finish, and slow the leasing pace to see which mixes still lead.

The point

Every rectangle on a floor plate is a bet. The mix decides how many bets you make, the corridor decides how many you lose before you start, and the stack decides when they pay out.

Draw the mix like it's money, because it is.

Test your mix, corridor and core position against your own rents before you lock the program. Book a TestFit demo and run this on your own site.