Planning a new house

Which market you are actually in

There are three ways a new single-family house comes into existence in the United States, and the Census Bureau counts them separately because they behave differently. A house built for sale is one a builder puts up on land it owns and sells finished. A contractor-built house is one built for owner occupancy on the owner's land, under the supervision of a single general contractor. An owner-built house is the same, except the owner acts as the general contractor. [1]

The proportions are lopsided. Of the 1,005,000 single-family houses completed in 2025, 721,000 were built for sale, 130,000 were contractor-built and 50,000 were owner-built. [2]

That matters because almost all published advice, and almost all published price data, describes the first category. If you are commissioning a house, roughly seven-eighths of what you read is about a different transaction: one where the builder chose the land, chose the plan, carried the risk and priced a finished product. Your project has none of those properties.

The order that works

Land, then constraints, then budget, then design, then permit, then build. That sequence is unglamorous and almost nobody follows it.

The common order is design first, because design is the enjoyable part and because it is the only part you can do before spending money. The problem is that a design produced without the site constraints is a wish, and every constraint discovered afterwards costs a redesign. Setbacks change the footprint. Soil changes the foundation. The height limit changes the roof. Utility distance changes the budget. Each of those is cheap to accommodate at the start and expensive to accommodate later, and they are all knowable in advance.

The single most useful discipline in the whole process is refusing to draw anything until you can write down, on one page, what the site permits.

What each stage locks

Every stage forecloses something, and knowing what is worth more than knowing what each stage produces.

Buying the land locks the setbacks, the height limit, the servicing distance, the soil and the flood exposure. Fixing the budget locks the achievable scope, which is area multiplied by specification and there is no third variable. Completing the design locks the structure and the mechanical strategy. Submitting the permit locks the documents, and any change after that is a resubmission. Ordering materials or modules locks the geometry.

After the last of those, the project has no cheap changes left in it. Everything from that point is a change order, priced by somebody who knows you have no alternative.

None of this argues for deciding everything at the start, which is impossible. It argues for knowing which door you are closing each time you take a step, so the decisions you defer are the ones that are genuinely cheap to defer. A kitchen layout can wait. A foundation type cannot wait past the point where the drawings go in for review, and a site constraint cannot wait past the point where you buy the land.

What this series covers

Four sheets, in the order the decisions arrive. Costs first, because the budget governs everything after it. Then the land and what it will tell you before you own it. Then permits and inspections, which are the schedule rather than an obstacle to it. Then the builder and the contract, which is where risk gets allocated between two parties who will not read it again until something goes wrong.

None of them gives you a price. That is deliberate and the reason is set out on the costs sheet. What they give you is the method for producing your own numbers, which is the only kind that will be right. The Build Sheet applies the same rule to every figure it publishes.