So, I'm in the position of considering where (within a given city/real-estate-market) I and my family want to build our next house (which will also be our first experience as homeowners -- we've lived in houses for some time now, but always as tenants). So far, we know that your typical suburban (new plat/subdivision) lot is out for three reasons:

1. Our current market is not exactly saturated with transit service, and most of what exists is constrained to more urban areas by the realities of transit efficiency. (Our current location is near the boundary of transit availability as is -- if we went any further out into the suburbs, we'd find ourselves in a transit desert)
2. I, personally, have no wish to contribute to sprawl by being a unit of demand for suburban real estate.
3. Builder attached lots are of no use to me (for reasons having to do with the hard/vertical costs discussed below), nor are HOA'ed lots (an HOA on a SFR has no upside and significant downside for me)

Furthermore, while I'm not expecting hugely out-of-line soft costs in my buildout plan, as the only major soft-cost adder item I foresee over a "typical" build is extra architect/engineer time, I am estimating fairly high vertical/hard costs (about 1.5x nominal is my conservative guess), and in particular, high hard costs devoted to structure, envelope, and systems in order to meet high performance and maintainability/durability requirements. This is somewhat counterbalanced in my mind by the lack of any need for a particularly large house (3000 ft² is large for me, if you want an idea, as our current situation is roughly half that size) and a preference for relatively modest finishes, amenities, and fixtures, but those also run contrary to the more normal equation found in typical high-end housing, where the structure, envelope, and systems are modest improvements over Code minima while the finishes, amenities, and fixtures are high-end and the house is very large.

However, all of this isn't a problem until you add the final ingredient into the equation: overimprovement risk. Market lots of equivalent size in established neighborhoods in my current market go for anywhere from 1/5 to 1/10 of the price of their suburban (new plat) counterparts unless there's something countervailing about the neighborhood, such as being in a historic area or a different school district from the primary urban district in the market. (There's one market lot in my target area closer to 1/3rd of the price of the suburban lots, but it's twice the size and carries that size as surplus land due to access issues.)

Given that houses in the vicinity of those market lots sit at a 9:1 improvement-to-land-value ratio by the assessor's numbers and have roughly half the valuation needed to breakeven at the forecast hard costs I've been able to come up with so far, how much risk am I taking on of overimproving an urban lot with my plans? Furthermore, are there strategies, such as facilitating future "house hacking" or strategically picking a location within the market that has countervailing factors upvaluing the land there, that I should be looking to apply to mitigate the risk that the house can't be sold when the time comes without taking the mortgage "underwater" with it?

For comparison purposes, BTW: new suburban housing stock in the neighborhoods with lots available carries a roughly 5:1 to 6:1 improvement-to-land-value ratio according to the assessor's numbers and what limited market data I could find, while existing stock in the aforementioned historic district and similarly situated adjoining neighborhoods sits around 4:1 to 6:1 as well. As to said alternative school district? Well, the properties near the one market lot available there are valued anywhere from 5:1 (interior lots of the same size as the subject lot) to 9:1 (corner lots) by the assessor, and the one nearby house for sale sits at about 2.5:1 in the market and 7:1 by the assessor on an identical lot to the subject.