Floor plate depth is the conversion variable I see mispriced more often than any other single line item.

The rentable yield framing is the right one, and the $150 versus $200 per leasable foot comparison makes the mispricing legible in a way that per-gross-foot underwriting does not.

The assumption doing the most work in the atrium scenario is structural system, as you note. Post-tensioned plates are the most common obstacle, but the beam transfer required when you interrupt a conventionally reinforced grid across multiple floors can close the gap faster than people expect. The structural engineer's preliminary opinion on opening feasibility, including whether a transfer beam can be hung without reconfiguring the lateral system, needs to come before the atrium is treated as a solution in the proforma rather than a cost variable.

The risk you did not mention is MEP routing on deep plate buildings. A 110-foot plate that was originally core-served for open office typically runs mechanical, electrical, and plumbing from a central core outward. Converting to residential means you need individual unit stacks, and if the core is positioned for office layouts, the routing distances to perimeter units on a deep plate can add cost that does not show up until rough mechanical drawings exist. On some configurations the rerouting eats a meaningful portion of whatever savings were assumed from ceiling height or existing infrastructure.

On the zoning question around interior non-windowable space, a few jurisdictions allow a secondary sleeping or living area without direct natural light if a light well or borrowed light arrangement meets minimum standards, but that is a local code and building official conversation, confirmed in writing before it enters your unit mix assumptions.

What is the structural system on the building you are looking at, and has a structural engineer done even a preliminary walk on the plate opening question?

Ran into exactly this on a small deal I almost partnered into in Lockhart, 2019, gutted two-story commercial block, previous owner had already stripped it to the slab. Plate ran 94 feet and the guy pricing it was quoting hard costs on gross area like they were all equal square feet. I kept asking him to show me the unit layouts and he kept showing me these bedroom closets that were 11 feet deep and facing an interior wall. That was not a closet, that was a room he could not figure out what to do with. We passed. Building sold again in 2022 and from what I heard the numbers still did not work.

Post-tensioned floors, though, that line about cutting through PT slabs being "a separate budget conversation" is being polite about it. I have sweated pipes through PT decks on commercial repipes and it is not a conversation, it is a different project. You cannot just core where you want to, you have to scan first, pray the tendons are where the drawings say, and work around them one at a time. An atrium cut through three or four PT levels is structural engineering, demolition sequencing, and temporary shoring before a single finish trade touches anything. If somebody has already priced an atrium into a conversion budget as a line item and the existing structure is PT, what number did they put on that line?

The post-tensioned slab point is where I want to add something, because the structural cost of an atrium cut is only part of the exposure. If the building financed its original construction with a CMBS loan that has since been defeased, or if there is any surviving covenant in the original deed or a historic tax credit regulatory agreement, a structural modification of that scale can trigger a loan assumption review or a credit recapture event even when the current owner thinks the capital stack is clean. That gets missed in pre-acquisition due diligence constantly, because the title search finds liens and the phase one finds environmental flags, but nobody pulls the original financing documents to check whether a physical modification covenant survived the last sale. On the interior footage problem itself, some sponsors price the dead square footage as amenity space, which the proforma accepts as a cost offset but which rarely underwrites to the same per-foot value as residential net rentable area when a lender is sizing a construction loan against stabilized NOI. The spread between what amenity space contributes to value and what it costs to build is the number that actually decides whether the atrium solution pencils versus just accepting the lower yield and adjusting the acquisition price accordingly. What does the existing capital stack on this building look like, and has anyone pulled the original financing covenants yet?

The piece that often gets underpriced alongside the atrium cut is the mechanical shaft relocation. Office buildings stack their HVAC and electrical risers through the interior core precisely because that dark center space is cheap to dedicate to building systems. When you cut an atrium or try to carve leasable area from that band, those risers have to move to the perimeter or get consolidated, and on a multi-story building that is a scope item that shows up late in the estimate and lands hard. Confirm with your MEP engineer what is in the core before you run any yield numbers, because the shaft footprint eats into that interior square footage before you even start the atrium conversation. I would also ask your jurisdiction whether interior units without operable windows satisfy residential ventilation code through mechanical means alone, because some cities allow it and some do not, and that answer changes what your 110-foot plate can legally produce. What is the original use of the building, and do you know yet whether the floor plates are post-tensioned or conventionally framed?

The variable I would add to plate depth is the relationship between plate depth and the building's original HVAC spine, because in a lot of mid-century office stock the mechanical shafts run through exactly the interior band that becomes dead square footage in a residential conversion. If those shafts have to stay for structural or code reasons, your usable depth shrinks further than the floor plate number alone suggests, and the cost to reroute or abandon them comes out of the same hard cost pool you are already watching thin out. A 110-foot plate with a centered mechanical core that cannot be relocated is a worse starting condition than a 110-foot plate with perimeter or distributed mechanical, even though they look identical on a depth-only analysis. Before accepting any rentable yield estimate I would want to see the existing mechanical drawings alongside the floor plate dimensions, confirmed by an MEP engineer who has actually walked the building, not just reviewed the as-builts. What does the existing mechanical layout look like on the floors you are planning to convert?

The 58-to-62 percent yield number is right but I'd push back on framing the atrium as a fix people "reach for" like it's a reasonable second option. On a deep plate office I looked at in Columbus last year, the atrium math killed the deal before structural ever came up. You lose the interior units, you add a new envelope to waterproof and maintain, and your HVAC zones get complicated in ways the first-round MEP estimate never captures. By the time we repriced it honestly the project needed rents the submarket wouldn't support.

What I'd do instead is walk away from the deep plate building unless the acquisition price already reflects the yield penalty in full, meaning the seller has already taken the hit that your construction budget will eventually force on you anyway. In Cincinnati I underwrote a 96-foot plate conversion last spring and the only way to make it work was getting the basis low enough that even at 63 percent yield the hard cost per leasable foot stayed under $185. We didn't get there. Seller thought the atrium option was an asset, not a liability. What plate depth are you actually working with, and what did you pay per gross foot?

120 a foot on gross is optimistic for anything priced in the PNW right now, we're running 155 to 165 before structural unknowns touch the estimate. That shifts both your endpoints up, so the spread between the 80 percent yield scenario and the 60 percent one is still real, but neither number is as comfortable as it reads on paper.

What's the vintage on the building?

Municipals are sleeping on this in Cleveland right now. The city is dangling adaptive reuse incentives for the downtown office stock and every pro forma I have seen floating around is using gross square footage as the denominator for return projections, not leasable. That math looks fine on a slide deck and then a 105-foot plate building goes under LOI and nobody catches it until the test fit comes back looking like a bowling alley with two dark rooms in the middle.

What I have not seen anyone price correctly is the carrying cost during the atrium permitting fight. If you are in a jurisdiction that requires a variance or a special use determination before the atrium satisfies egress, that is four to eight months of holding costs on top of whatever the structural work adds, and that window gets worse if the building is in a historic district.

What jurisdiction are you working in, and is the plate depth uniform across all floors or does it widen at the podium?

My cabin conversion quote last spring came back identical per gross foot whether the interior footage was usable or not, and the contractor never mentioned it.

Floor plate depth prices the deal, but the number that actually sets your cost basis is the ratio of perimeter linear feet to gross floor area, not plate depth alone. A rectangular 110-foot plate on a narrow building reads completely differently from a square 110-foot plate, because the square gives you a much worse perimeter-to-area ratio and a larger dead interior band relative to the rentable shell. Two buildings with identical plate depth can have a $40 or $50 per leasable foot spread between them just from footprint geometry, before a single design decision gets made. What is the building's aspect ratio, and have you already done the perimeter unit count at a standard 12-foot module to see how many units the geometry actually supports before you price the atrium option?

Atrium cuts always pencil on paper and then the PT slab bid comes in.