A lumber package repriced 9 percent between two spec houses on the same street, and here is how a builder might think through it

85k gross on 15 months with a 460k basis is roughly 18 percent on cost before your own time, and it's before any price concession, any commission if you use one, and any overrun. I hit this same ceiling last year on my third project and what I learned is that the margin that looks acceptable in the model is the margin that goes to zero the first time something real happens.

I'd start B only if I could get the basis under 440. That means either the lot was cheaper than 88 (it wasn't, you already own it) or the build comes down. Can you value engineer the mirrored plan? Same footprint, cheaper roof geometry, drop a bump-out. On my last one I took 9k out of the hard cost by simplifying the roof and nobody who walked through it noticed.

Coming from zero here so tell me if this is a dumb read, but the 9 percent framing reprice, is that a quote you can still lock? Or is it a moving number you'd be exposed to all the way through framing? I'd think whether you can fix that price changes the whole decision.

@bmoreno23 that is a sharp read, and it is the question to get answered first. @carlos_nunez, if your supplier will hold that package price for 60 days in writing, House B's budget has a floor. If he will not, your 338 is a guess and the real number could be 350. Separately, the builder three streets over cutting 15k and still sitting is the data point to weight highest. That is the market telling you 545 has a soft ceiling, and it has nothing to do with your costs. Two unsold houses in a submarket where the last comparable seller had to discount sounds like the thing you would write a post about in November regretting.

What does your absorption actually look like? Set last summer's four closings aside and count the active listings in that price band right now against how many closed in the last 90 days. If there are 11 active and 3 closed a quarter, there is a year of standing inventory ahead of you and House B is a bet on 2027 buyers. The other thing worth knowing is your personal liquidity outside the projects. The figure itself matters less than whether you can carry both for six months past plan without selling something. That question decides this more than the margin does.

One option nobody has raised. If House A sits, what does it rent for? In some markets builders have gone ahead and leased finished specs at a loss to interest to stop the bleeding while they wait. A 545k house probably rents for 2,900 to 3,300 depending on the market, and your interest alone on 340k at 9.75 is about 2,760 a month. So renting roughly covers debt service and nothing else, and now you have a lease encumbering your sale and a tenant-occupied showing problem. It is a real exit, it is just a bad one. Worth pricing so you know what your floor looks like before you commit to B.

That is useful even as a bad exit, @tonyv88, because it turns "how long can I sit" from a feeling into a number.

The framing here is A or B, and there is a third option: start B on a stretched schedule. Break ground, get the foundation and rough in done, then slow walk the finishes until A closes. You hold the permit and keep the framer engaged for a few weeks, and you do not commit the expensive half of the budget (cabinets, flooring, fixtures, appliances) until you have a data point from A. The cost is carry on partial draws and a house sitting half done, which invites weather damage and theft. It still beats a binary.

@luis_carrera building on that schedule is more expensive than it sounds. You demobilize and remobilize every trade, and each restart costs you a slot in your subs' queue. When 94 percent of contractors say they cannot fill positions, the sub who moves on to another job comes back when his own schedule allows, which is rarely when yours needs him. A two week pause routinely turns into seven. @carlos_nunez, the concrete number moving from 11.40 to 12.60 because he lost two guys is the tell in your whole post. That is a labor supply problem rather than a materials cycle, and it is structural. Price your next 24 months assuming subs get more expensive. On value engineering, @samantha_ruiz is right about roof geometry. Check as well what your plan spends on window count and header spans. Mirrored plans often end up with a worse orientation and people add glass to fix it.

The second project is where an operator finds out whether the first one was skill. You have 218k in the ground on A and no closing yet. You do not know your real cost to complete, and you have no days on market or closing table net to measure against. Starting B before A closes means scaling a process you have not measured once. Sit on the lot and eat the 640 a month, and pay whatever the permit extension costs. That is maybe 8 or 9k of dead money over a year. Cheap tuition compared to two unsold houses.

The permit clock is usually less fatal than people think, though it varies by state and by local jurisdiction, so call your building department and ask exactly what keeps it alive and what an extension costs, in writing. In many places a permit stays valid as long as work is progressing on some interval, and an inspection on a footing can restart the clock. Some jurisdictions charge almost nothing to extend. Holding a paid for lot next to your own completed spec is a decent position. You become the one seller in the submarket who can watch A's absorption and then start B with real pricing information. Land is patient in a way a framed house is not.