Why I Build for Niches Silicon Valley Ignores

The three projects on my desk right now look, on the surface, like they have nothing in common.

Jettison Air is a command center for private aviation charter brokers. Roof Guru is a multi-tenant platform for residential roofing contractors in Nashville. League-OS runs a youth flag football league for 1,500 families and just shipped a check-processing pipeline that spans Vercel and Google Cloud.

Private jets. Asphalt shingles. Third-grade quarterbacks.

Look a little closer and there is a thread. Each one serves a niche the big SaaS players refuse to go near. Each one was built from inside the operation, not from a deck. And each one is defensible for the same underlying reason: I went deep where nobody else bothered.

The Generic SaaS Playbook Is Exhausted

For the last ten years, the default move for any builder with a laptop was to pick a horizontal problem, wrap a clean UI around it, and sell it to as many industries as possible. CRMs. Form builders. Scheduling apps. Project management tools. Anything that could plausibly work for "small businesses" got funded, launched, and eventually buried by whoever had the biggest ad budget.

That playbook is a red ocean now. AI drops the cost of building the generic thing to near zero, which means the only surviving advantage is distribution, and distribution is a game of money and momentum that a one-person studio is never going to win.

What is still rare, and still valuable, is depth. Depth in a specific vertical. Depth in a specific operation. Depth that took years to acquire and cannot be generated by a prompt.

Why Niches Actually Win for a Solo Builder

Four things hold up when I pressure-test this against real work.

Operators pay for pain relief, not features. A roofing contractor does not care about a beautiful Kanban board. They care about whether your estimate calculator knows what a waste factor is. A charter broker does not need more integrations. They need the branded PDF to render correctly every single time, because that document is the product. When you serve a niche, the feature list gets shorter and the value gets sharper.

Competition is shallow. The number of builders willing to spend six months learning the specifics of FAA Part 135 operator certifications is small. The number willing to learn ACV versus RCV insurance pricing is smaller. The number willing to run a flag football league for three seasons before writing the software for it rounds to zero. Nobody is competing with you on depth, because depth is expensive.

Trust compounds. One operator in a niche talks to every other operator in that niche. Reputation travels on foot at conferences, in Facebook groups, in the text threads that happen after a Tuesday night league meeting. Shipping one real thing for one real operator opens more doors in a vertical than any amount of paid acquisition.

Support load stays bounded. When your customer base is people doing the same job in the same industry, their questions converge. You get to build a playbook, not a call center.

Three Ways I Pick a Niche

I have a simple test. I will only build for a niche if one of these is true.

I operate in it. League-OS exists because I actually run the league. Every feature came out of a real Tuesday night crisis with a real parent on the phone. The Field page on this site is the short list of businesses I personally run or co-run, and it is the top of the funnel for almost every product I ship.

I partner with an operator. The landscaping business is a partnership. I handle the back-office software, my partner handles the sales and the industry knowledge. The only way I learn how the trade actually works is by being close enough to someone who lives in it to see the seams.

A client brings me in deep. Jettison Air started this way. Roof Guru started this way. A specific operator with a specific, underserved workflow hires me to build the tool they wish existed. The product is the result of months of watching their week, not a spec document.

If a niche fails all three tests, I do not build for it. That is a rule, not a preference.

What Depth Actually Looks Like

The easiest way to show what niche depth buys you is with a concrete example. League-OS has to let coaches post payments for cash, check, and money order, which is still how most youth sports payments happen. That sounds boring until you try to actually process a photo of a check.

Here is what it takes to do that well, in a Vercel plus Google Cloud setup:

1. The browser requests a signed upload URL and writes the check PDF straight to Google Cloud Storage. No proxying through the app server.
2. The Vercel app calls a containerized service on Google Cloud Run, authenticated with an ID token minted by the Google Auth Library.
3. Cloud Run reads the PDF, renders every page to a PNG using pdf-to-img and @napi-rs/canvas, and writes the images back into GCS.
4. A job engine in Postgres orchestrates the multi-step work. Each step runs inside a Vercel route and uses after() to self-chain to the next step, which keeps every invocation under the 60 second serverless cap.
5. Lease-based serialization prevents two workers from processing the same job. A cron recovery job picks up anything stale.

None of that is generic SaaS plumbing. It exists because one specific niche, youth sports, still runs on paper checks, and nothing off the shelf handles that cleanly. You cannot build this out of features. You can only build it out of understanding how a real league actually collects payments on a Saturday morning.

The Short List

If you want the map of where my work is pointed today, it lives on the new Industries page. The short version:

• Private aviation, with Jettison Air.
• Residential roofing, with Roof Guru.
• Youth sports and recreation, with League-OS and AFL Photos.
• Faith communities, with Biblical Battle Plans.
• Landscaping and field services, from the Field.
• Health and self-tracking, with the Personal Health Engine.

Every one of those is a vertical I either operate in, partner into, or got pulled into deeply by a client. That is not a coincidence. It is the filter.

If you are running a business in a weird, specific, unglamorous niche and the software you have been handed does not fit, I want to hear about it. The weirder, the better.

Relentlessly curious. Unreasonably willing to build.