Bent Tree Valley, Part 1: Why an Engineer Builds His Own Dream House

Land, Lessons, and a Shop First — what two sets of plans, one rejected quote, and a lot of risk management taught me.

House

It started on our honeymoon, nearly a decade ago. Meghan told me she'd always wanted to build a house in the country. My dad did exactly that when I was five, and I'd wanted the same ever since. I'd owned homes for 13 years and built a shop at one of them, but building a house from nothing was a different scale of intimidating. Since we both wanted it, the dream became much more real, so we decided it was something to start working towards.  The first step would be land. 

On that honeymoon, Meghan found it: 40 acres, half timber, with a spring that becomes a waterfall after enough rain. I had enough for a down payment, so we pulled the trigger. Three months after the wedding, we owned 40 acres of virgin ground.

Attempt #1: Hire It Out

My plan was simple: buy house plans, hire a general contractor, and let them handle just about everything. Meghan was pregnant by then, so I didn’t want to get spread too thin.

Then I ran into the market as it existed in 2018. I'd call it a job shop vs. production line problem. Production housing, the neighborhood with an HOA, is a genuinely efficient system. Standard floorplans, shared infrastructure, and predictable resale make the unit cost low and the process repeatable. It's the right answer for a lot of families, especially those who move for work every few years.

But I'd learned from buying my last house in Kansas that I wanted something a production line doesn't offer: room for a shop, a place to park my school bus, and the freedom to do what I wanted with my own property. That meant a one-off build, and one-off builds are job-shop work. Every cost that a neighborhood spreads across hundreds of lots lands on one: the driveway, the electrical service, the well and septic. On raw land, that was a $50,000 setup cost before we'd framed a single wall.

I found a half dozen contractors, and exactly one who answered. He was good with large remodels but new to ground-up construction, and his first quote came in 50% over budget. We redesigned, paid for a second set of plans and an engineer's stamp, and got within 10%. My banker, who saw a lot of projects, noted that everything he funded was running 10-15% over. The appraisal compounded it: with so few comparable rural homes, there was barely any data to support the value. Our budget was hard-capped, a baby was on the way, and our equity was on the line.

So I walked away, about $5,000 poorer for two sets of plans and an appraisal. It was a painful result but the right decision. Too much risk, not enough data.

Attempt #2: Manage the Risk Myself

Walking away left me with a more useful question: how did my dad do it in the '80s? He bought a kit, worked with the bank, and did much of the work himself. About 20% of that failed quote was contingency, essentially the contractor's price for uncertainty. If I took on the uncertainty myself and managed it down, I could keep that money.

That's a risk management problem, and risk management was already my day job. Between my Army background and years in quality engineering, regulation and project risk weren't foreign. I'd even contracted development projects in Afghanistan. I just hadn't wanted to do it for myself before.

It turned out house plans are far less mysterious than advertised. In our first attempt we bought a set for $1,000, marked them up beyond recognition, and the engineer stamped them without drama. The same production vs. job shop dilemma existed with the plans:  Each minor change cost $500, and none of the off-the-shelf designs were just a change or two away from what we really wanted.  I missed doing CAD work anyway, so it seemed the perfect time to take on architecture as a hobby.

The real unknowns were contracting, permitting, and building in our county, and I wanted to learn those on something smaller than the house.

Shop First

The shop was the perfect pilot project. I'd outgrown a standard 2 or even 3 car garage years earlier and needed room for a fabrication bench, drill press, hydraulic press, welder, and lathe. (Funny how a wrenching hobby turns into a construction hobby.) We looked at a barndominium and found it really wasn't any cheaper than building a shop and house separately, so I built the shop first at roughly one-sixth the scope of the house. It let me practice electrical, plumbing, framing, and about half of the site development, while a remodel of our current house covered the finish carpentry.

I broke ground in 2020, just after our daughter Ophelia's first birthday. Permitting, inspections, and county requirements taught me plenty, and I was happy to fail fast on a project where mistakes were cheap. I closed the shop permit in early 2022. A promotion followed, which funded the house but cut into my time, and our youngest, Genevieve, was born that summer.

Where We Are Now

A few things have changed. Several new local builders have appeared since the pandemic, and services like U-Build-It now absorb some of the biggest DIY risks. The market gap that stalled me has largely healed. Launching Big W set the schedule back again, but we're finally close to ready for the next step.

Meanwhile, I've spent six years refining the house design, about 180 revisions across two CAD platforms. The last two years brought only one or two per year, so the model is mature. The land is taking shape too: an orchard with apple, peach, and cherry trees, and a privacy treeline.

What's Next

In upcoming installments I'll share how I'm applying Six Sigma tools to a house build:

• Design requirements: using a Pugh matrix to align Meghan's and my priorities before drawing a single wall

• Cost control: building a digital twin time-and-cost model to find both savings and put a dollar value on upgrades

• Shop and orchard: what went well and what I'd do differently

Bent Tree Valley is my favorite real-world lab for the idea behind Big W: with the right systems, you can build almost anything from scratch.