What a Third Garage Taught Me About Plan Data

Over my years in architecture and product leadership at builders like Pulte and Beazer, I kept running into the same problem. We’ve modernized nearly every part of the business, except the one thing everything else depends on: the homes themselves.

This gap has a real, measurable cost. It shows up as margins we can't fully explain, land and product decisions that take months instead of days, and a company that knows it could compete harder if it could just see what it already owns.

For months, I couldn't figure out why our Capstone plan cost more to build in Minnesota than in Indiana. Same drawings. Same square footage. Same everything, on paper. My counterpart in purchasing and I went line by line through drywall, concrete, framing, hunting for where the money was going. It took us months of manual comparison before we found it: one division had quietly made the third-car garage standard in their market.

The answer had always been in the home itself—we just couldn’t see it. We had no way to understand why there was a discrepancy or tie the actual costs back to the product. If the home had existed as structured data instead of disconnected drawings, the answer would have taken minutes, not months.

We know everything about our customers. We know everything about our transactions. Our CRM and ERP systems make sure of that. What we've never captured with the same rigor is the plan itself, the actual home we're building.

Standardization doesn’t solve the problem

Every time a builder has tried to operate more efficiently by standardizing on fewer plans and limiting options, they’ve run into the same reality: home building is local. Vendors, materials, jurisdictions and buyer expectations differ market to market, and local leaders are accountable for making their numbers work under those conditions. That's not going away, and it shouldn't.

That division in Minnesota may well have made the right call: a buyer expectation nobody could ignore, a builder association standard, something the local team understood better than corporate ever could. What actually hurt us is that the decision disappeared the moment it was made. There was no record anywhere that would let someone outside that division see it, learn from it or account for it in a cost comparison.

I saw the same pattern show up on the purchasing side. Divisions that pay trade partners a flat rate, say $5,000 for a finished bathroom, often have no idea what's actually inside that number. One sink or two? A tub or just a shower? Even within a single division building a handful of communities, that adds up to real money you can't see or negotiate against, because nobody can say with certainty how many sinks or tubs actually went into the homes you built last year.

The most valuable data is hidden

The most valuable data in a builder's business is the data nobody can see: design decisions buried in drawings, tribal knowledge, checklists and spreadsheets that only make sense to the person who made them.

This is the problem we've built Higharc around solving, and it's a different approach than just digitizing paper. A scanned PDF of a floor plan is still just a picture. That picture doesn't know what a wall is, what a cased opening is or what separates a bedroom from a closet. Instead of trying to pull data out of 2D pictures after the fact, we create the data first, then use it to produce the drawings and other assets home building actually runs on. Every wall, room and fixture is an actual, defined piece of data rather than a line on a page.

That structure lets me ask real questions of my plan library instead of just storing it. At a national builder, that meant looking across dozens of divisions. But the same blind spot shows up at a builder running two or three communities in a single market: a "custom" option in one community turns out to be the same product already built somewhere else, or two plans that were supposed to be different are actually within a few inches of being identical. You're not going to catch that in a stack of drawings or a shared drive full of CAD files. It requires structured data to compare, at any scale.

It also means when a land team asks whether we have a product that fits a specific lot, I can actually query the answer instead of guessing or starting from scratch.

If my homes were data, I would have saved months chasing the Minnesota and Indiana cost gap. With Higharc, I can automatically pull up cost deltas by room, by wall or any comparison I want. Within a minute, I could have looked at the Minnesota plan and seen we were paying more specifically because of the garage area built into their base plan.

AI won't work across the business without the plan data

Having this level of product data also opens up new AI applications, because the AI can finally reason about a home the way I do. It already knows whether a line on the page is a support wall or a decorative one, because we defined that distinction when we built the plan in Higharc. That's what lets AI start analyzing our plans at a scale no person could, in a way that actually informs business strategy instead of just automating busywork.

That's also a big part of why our recent $95 million Series C matters to me personally: it's going directly into deepening our data model, so our AI gets more capable over time.

When homes become data, everything changes

Once plans exist as data, we can manage our product strategy with the same rigor as our customer relationships and financial performance. Land and product decisions that used to take months can take days. Sourcing conversations can be backed by real volume instead of estimates. And the operators who already know their markets get real evidence to work from, instead of relying on memory and guesswork.

None of that requires giving up local flexibility. It just requires finally being able to see it.

Learn more about how Higharc turns drawings, tribal knowledge, checklists, and spreadsheets into structured product data that helps builders improve time to market and margins.

Links to sponsored content are being provided solely for the convenience of 20 Club members, and for information purposes only. Providing links to sponsored content does not imply NAHB’s endorsement, recommendation or approval of any content provided by sponsors and hosted or maintained on sponsor websites.