The setup
A custom cabin builder sells a complicated product. The lineup launched with nine models, and each one comes with dozens of options: size, layout, exterior, interior, add-ons.
The way they sold it, before we got involved, was the way most custom builders sell: by hand. A customer filled out a contact form. A rep took those details into a separate tool and built a proposal. The proposal went out by email for a signature. Then somebody followed up on the payment by hand. Every step needed a person. Every step added delay, and every delay is a chance for a buyer to cool off, call a competitor, or just get busy and forget.
The product wasn't the problem. The process between "I'm interested" and "here's your price" was the problem.
What we set out to build
One place where a buyer could build their cabin and see a real price, for the whole lineup, without waiting on anyone. Not a brochure with a form at the bottom. A working pricing engine.
And we designed it as more than a website feature. A configurator like this is the front end of the entire revenue process: configure, propose, sign, pay. Build the front door right and the rest of the house can be wired to it.
What we built
At launch, the configurator covered the builder's full lineup: all nine models, each with its full set of options, priced in real time. Change a size, change a layout, add an option, and the price updates as you go, from one pricing engine backed by a managed database. No "request a quote." No waiting a day or two for a number. The buyer does the exploring, and the number is always there.
Dealer pricing sits inside the same system, gated behind a password, so the people who should see it can and the public can't. Access is controlled at the level of the data itself, not just hidden behind a page.
The platform it runs on costs about $20 a month in infrastructure. That's worth pausing on. This is not a six-figure software suite. It's a well-built system on modern infrastructure, owned by the client, and it runs their entire public pricing.
It also wasn't built as a dead end. The architecture layer that lets AI tools work with the business's data went in at build time, so when a better model comes along, or the builder wants to run one privately, the switch is a single change in one place. The system stays. The model is replaceable.
What we measured
The configurator has been live since April 2026, with the full lineup priced in real time. That's the headline measurement, and it's the kind we like best: not a projection, not a launch-day screenshot, a system still running and still selling months later.
The downstream pipeline was designed end to end at the same time, as the next phase: a submission would generate the personalized proposal, the proposal would go for electronic signature, and the moment it's signed, the deposit link would fire and the rep would get notified.
The payment design would change the math on every sale. The old client-management platform took a flat 1.5% of every payment. The new design would route payments through ACH at 0.8%, capped at $5 per transaction. On high-ticket cabins, that cap is the whole story: the fee stops growing with the price.
What surprised us
The configurator looked like a website project. It turned out to be a sales process project. The hard part was never the screens. It was the data: every model, every option, every combination, priced correctly, in one place, kept current. Get that foundation right and the website almost builds itself. Get it wrong and no amount of design saves you.
That's the lesson that keeps repeating in our work. Owners ask for the visible thing, the tool, the automation, the AI. The visible thing is the last 10 percent. The first 90 is the unglamorous foundation underneath it, systems and data that agree with each other.
What an owner should take from this
If your team rebuilds the same quote by hand every week, you don't have a quoting task. You have a system that hasn't been built yet. The hours are visible if you look: the form, the second tool, the email, the follow-up. Add them up over a year and the manual process is usually the most expensive software you own, and you don't even get a login for it.
Second, buyers move at the speed of answers. A customer pricing a cabin at 9 p.m. is a customer. If the price depends on a rep's Tuesday, some of those customers won't be there Tuesday.
Third, own what you build on. The builder owns this system and the data in it. The tools underneath can be swapped. That's the difference between a system that's an asset and a subscription that's a landlord.
We stay on to run what we build, and this one has been running since April. That's the whole pitch, really: build the thing that sells, on a foundation that lasts, and keep it running.