B2B SaaS · 0→1 · Designed & built end to end
Solobilt
Websites for contractors who went solo. Built for speed to market: I designed and built it end to end to get a real product in front of contractors fast.
At a glance
Problem
Solo tradespeople need a site that earns a homeowner's trust in seconds. What they have is a blank Wix theme they don't have time to fill, or no site at all, and generic builders bury license, insurance, and years on the job under “About Us.”
Solution
Paste a Thumbtack, Google, or Yelp link and see a trade-specific site in about 10 seconds, with no account needed. License, insurance, and service area sit in the hero, and distinct theme families replace color swaps.
- Paste a linkThumbtackGoogleYelp
- Site in about 10 secondsTrade-specific, no account needed
- Trust in the heroLicenseInsuranceService area
- Theme familiesDistinct families, not color swaps
Impact
- Designed and built end to end by me
- Live in pilot at solobilt.vercel.app
- Pilot test: I ran a paid ad that drove 200 clicks to the Home page

01 / Problem
Win a homeowner’s trust in seconds
Homeowners look a contractor up before they hire. The page that wins the job puts license, insurance, service area, and real job photos where they look first. Generic builders treat a plumber like an agency, so that proof ends up under “About Us.”
This came from interviews with 5 contractors.
02 / My role
One person, whole stack
I designed and built Solobilt end to end: positioning, onboarding, the site-generation theme system, pricing, and the marketing site. Working alone, every decision had to hold up in design and in code. Tools: Grok Bot to build, Vercel to ship.
Constraints
- Users won't fill out forms. A contractor between jobs has minutes.
- Data I don’t control: each profile source exposes different fields, or none.
- Trade-specific, not one template with a new color.
03 / Key decision
Start from a link, not a blank page
Paste a profile link and see your site in about 10 seconds, with no account. Claim it when it looks right. If you already have a site, the “Replace my site” path pulls its title, description, and contact details instead.

04 / Key decision
Proof above the fold, nothing faked
License number, insurance, years on the tools, and service area by town name sit in the hero strip, because those are what reassure a homeowner. Trust is the pitch, so anything that isn't real is labeled: example ratings say “example,” imported reviews say “imported.”

05 / Audit
Auditing my own output
I generated 8 example sites and audited them side by side. They were one design with color swaps: the same skeleton, type, grid, stock photos, and fill-in-the-blank copy. I benchmarked against Squarespace's template library, which varies hero structure, grid, type pairing, and section order, not just color.
06 / Key decision
Variation is the product
I replaced color swaps with distinct theme families, each with its own hero structure, type pairing, and section order. Trade-specific details sit on top, like drawn pipe icons, brick bond patterns, and a garden-year calendar.


07 / Pricing
Two doors, one product
Contractors can do it themselves (Basic or Pro, no setup fee) or have us build it (Done-for-you). Both use the same generator and dashboard, so done-for-you is just us doing the onboarding.


08 / Outcome
In pilot
Live in pilot at solobilt.vercel.app, with a dashboard where contractors update copy and photos when a job wraps.
What I'd do next: test the preview-to-claim step with real contractors, measure which theme families convert, and find which entry point brings them in: pasting their current site, the CTA buttons, or the pricing section.
Outcome
- 12 example sites across 11 trades, labeled on the site as fictional demos: plumbing, electrical, HVAC, landscaping, masonry, painting, roofing, general contracting, bath and kitchen remodeling, windows and doors, and handyman.
- Pilot test: I ran a paid ad that drove 200 clicks to the Home page

Next project
Prototype HandoffA concept where CAD filenames fill in prototype specs, so designers stop re-keying data. Concept, never built.