How a multifamily site gets chosen, parcel by parcel

Most of the ground that looks like an apartment site is not one. A developer walks many parcels for every one that gets built, and the ones that fail rarely fail because the corner was bad. They fail on a sewer line running the wrong direction, on a grade that eats the parking, on a drive the county will not permit where the plan needs one, or on arithmetic.

Choosing a market is a separate decision, made earlier and on different grounds; why we build in six Ohio markets covers that half of it. Site selection is what happens next, when someone stands on a specific piece of ground.

What a site screen is for

The screen’s job is to kill sites cheaply, and in the right order.

Site control usually begins as an option or a long diligence period rather than a purchase — the right to walk away while you learn what the ground can carry, which is where a development deal starts. The clock runs and the studies cost money, so the questions get asked worst first: the absolute ones, that cannot be designed around.

A parcel that cannot get sewer is not a site. It does not become one because the ring around it looks good, or the price is attractive.

Sanitary capacity, and where the line actually runs

Sewer usually goes first, because it has the fewest workarounds and the most expensive ones.

Two questions hide inside it. The first is physical: is there a sanitary main the site can reach, and can it reach it downhill? Gravity sewer runs one way. A parcel below the line needs a pump station and a force main — capital cost up front, then a maintenance obligation somebody owns permanently. If the main is not adjacent, the project pays to extend it, usually across ground belonging to people with no reason to grant an easement.

The second is capacity: whether the system downstream will accept what the project adds. That belongs to the sanitary district or the city engineer, and the answer can turn on a constraint miles away — a segment or a plant at its limit, with new connections restricted until it is relieved. Ask the utility in writing, early, rather than taking the seller’s account of it.

Water and power ask smaller versions of the same question.

Grade, and where the water goes

Topography becomes money twice.

First, earthwork. Buildings sit on flat pads, so ground with fall in it gets cut and filled until it has them. Balancing dirt inside the property line is manageable; importing it or hauling it away is expensive and slow. Steep ground also brings retaining walls, longer utility runs, and accessible routes that get harder to hold to the slope the code requires.

Second, storm water. Roofs and pavement replace absorbent ground, so water leaves faster than it did, and the jurisdiction requires it to leave no faster than before. That means holding it, in a basin or under the pavement. Detention competes with the buildings and the parking for the same acreage, which makes it a site-selection question rather than a civil detail for later.

Getting in and out

A parcel with no workable access is landlocked in every way that matters.

Start with who owns the road. A municipal street, a county road and a state route are three permitting authorities with three sets of standards: how far apart drives must be spaced, how much sight distance one needs, and whether left turns out are allowed. Those standards decide where a curb cut can go, and sometimes that a parcel with narrow frontage supports one drive, in the place the layout wanted a building.

Fire apparatus has to reach the buildings and turn around, and a long site on a single entrance can be required to provide a second point of access — easy when the adjoining ground is yours, a negotiation when it is not.

The traffic study and the formal confirmation of utility capacity belong to the approval rather than to the screen; where entitlement is decided sets that stage out. Here the question is narrower: can a drive physically and lawfully go where the plan needs one.

Floodplain, wetlands, and what is in the ground

Three record checks, run early because they are cheap and any one of them can end it.

• Flood. The FEMA map shows whether part of the parcel sits in a special flood hazard area. That rarely kills a site by itself, but it takes the affected ground out of the buildable area, and altering it pulls in a map revision on a federal timetable.
• Streams and wetlands. A delineation identifies which features are actually there and whether they are jurisdictional. Impacting one is not necessarily prohibited, but it requires permitting and mitigation — a cost, and more to the point a schedule set by an agency rather than by the project.
• Contamination. A Phase I environmental site assessment reviews the parcel’s history, the regulatory records, and the uses next door. A former filling station, a dry cleaner, or undocumented fill produces findings that send the work on to Phase II sampling.

Contamination is the question whose answer is most often to walk away, which is why it gets asked before serious money is at risk.

Whether the homes fit the shape of the parcel

A site plan is a packing problem. Buildings, drive aisles, parking at whatever ratio the code requires, setbacks, buffers where the neighbor is residential, the detention, the amenity building and the fire lanes all have to fit inside an irregular boundary, around easements nobody may build over.

Two parcels of identical acreage do not hold identical numbers of homes. A long narrow one spends more of itself on drive aisle; one crossed by a utility easement gives up a strip through its middle; one fronting a boundary the code buffers heavily loses a band along that edge.

The number of homes is what pays for everything else, so the shape of a parcel is an economic fact before it is a design one.

What the neighbors already look at

Two directions, and both are answered by driving the ring rather than reading a map.

• Outward. What residents will look at, hear, and live beside. A rail line, a loading dock, a highway embankment or a vacant strip center are facts about the leasing, not about the view.
• Inward. What the people already there will see when the building goes up. A parcel behind a row of single-family back yards is a different proposition from one across from a shopping center, whatever the zoning map permits in both.

The inward view does not decide whether a project is worth doing, but it predicts how the approval goes and how long it takes, and both are budget lines. The communities index lists ours with their addresses.

The arithmetic that rules out most of the rest

Everything above is physical. What eliminates most of the sites that survive it is a subtraction.

The findings of the screen are costs, not design details. A pump station, a retaining wall, extra detention and an off-site main are budget items in the way the framing is, and they are added to the land price before the arithmetic starts. Entry basis rather than rent growth is what a development argument rests on, and every one of those items raises the rent the plan requires while doing nothing to the rent the market will pay.

Then the test. Our pro formas use untrended rents, with assumptions set at market — the firm’s own words on how it underwrites — and a deal that does not make sense at today’s rents is not good enough for us to build. That is a claim about method and not about outcomes; what it changes once a building is leasing is a different question from what it does at this end.

What it does at this end is fix when a site dies. An untrended test can be run on the day the parcel is walked, because the rent it uses is one somebody is already paying. A trended one moves the answer into a year that has not arrived, so a marginal parcel keeps passing until the option expires and the money is spent. A rent test only rules a parcel out if it is applied while walking away is still free.

That is why most clean sites do not get built. A parcel can pass every physical test above and still cost more to develop than its own ring supports today, and the honest answer to that is to put it down.

What passing the screen does not settle

A parcel that clears every question above is a candidate, not a project.

Approval can be refused, or granted with conditions that change the building. Borings can contradict what the surface suggested. Bids can come back above the estimate, and construction costs and interest rates move between the day a site is screened and the day a loan closes. A screen run well makes those outcomes less likely; it removes none of them, and the chain that starts when one of them lands runs the same way however good the site was.

Development involves substantial risk. Private real estate is illiquid, no return is guaranteed, and investors may lose some or all of what they commit. Nothing here predicts how a particular site will turn out.

A screen can still be judged in advance: how a sponsor runs one, what it does with a parcel that fails it, and whether the people running the screen are the same ones who pay for what it missed. That last one is what vertical integration changes for an investor. To ask about a specific community, start with investing alongside us or put the question to our team.