How We Built Some Security Vendor (and Why)

Photo by Ales Nesetril on Unsplash.

Last week we shipped somesecurityvendor.com.

It's a fake security vendor landing page. AI-native. Shift-left. Developer-first. Built by practitioners. SOC 2 Type II. Everything you'd expect — because it's all been trained out of the same playbook.

This is how we built it, what we used, and what the reactions revealed.

---

The idea

The premise came from a real frustration buyers keep voicing: by the third demo of the week, they had stopped reading landing pages. The category had become a filter, not a signal.

We wanted to make that feeling visceral. Not an essay about it — a thing you click through and feel.

Build a perfect generic security vendor. Name, tagline, buzzword bingo, the full deck. Then reveal that you just described 200 real companies, not one fictional one.

---

The funnel

The site runs four stages, each doing different work:

1. Fake hero — A convincing but hollow landing page. AI-native. Shift-left. The words that stopped meaning anything.
2. The reveal — "This is SSV. Some Security Vendor. Also known as every other vendor on the conference floor."
3. The quiz — Five questions about the vendor's approach. Every answer is correct. That's the point.
4. The result — "It doesn't matter what you picked." Every answer described SSV — which is also DevShieldAI, CloudGuardX, TrustNativeCore, and 1,440 other names generated from the same prefix-suffix-modifier arrays the market has been using for a decade.

The name generator is PREFIXES × SUFFIXES × MODIFIERS: twelve prefixes, twelve suffixes, ten modifiers. The market has filled most of the combinations.

---

The build: v0 → Cursor → Vercel

Built in one weekend. Stack: Next.js App Router, Tailwind, Framer Motion, Vercel Analytics. The process:

v0 for the first hour. Started in v0.dev to rough out the hero section and quiz card component. Fast for getting from zero to clickable — the output needs work, but the blank canvas problem is gone in under an hour. That's the right use: eliminate the blank page, not write production code.

Cursor for everything after. Once the structure existed, I moved into Cursor and didn't leave. The entire four-stage funnel logic, animated transitions with Framer Motion, live counter, buzzword leaderboard — all written and refined there. Project-wide context is the unlock: it understands what you already built, so you're always moving forward instead of re-explaining.

Claude for the mechanics that aren't code. The quiz needed every answer to describe the same vendor without feeling like a trick. That's a framing problem before it's a code problem. Solving it in conversation is faster than solving it alone. Claude also helped pressure-test the copy — "does this land as observation or gotcha?" is not a question a linter answers.

Vercel for deployment and analytics. Stage-aware funnel events wired from day one: stage_view, primary_cta_click, outbound_click, each tagged with stage and destination. The KPIs we care about — reveal → Beyond Features CTR, result → newsletter CTR — are visible without additional infrastructure.

---

What the reactions revealed

The responses split into two camps almost immediately.

People in security and devtools GTM got quiet. Not "haha funny" — more like recognition. I wrote this deck. Some sent DMs. A few asked for the teardown.

Everyone else thought it was a real company.

That gap is the finding. When a satirical vendor page is indistinguishable from the real thing, the problem isn't the joke. It's that the category has trained buyers to expect one soundtrack — and the copy has obliged.

---

What I'd do differently

The quiz questions could be sharper. Five questions do the same job; one well-crafted question that collapses in the reveal would be cleaner.

And v0 is better used earlier — at the wireframe stage, not the component stage. I built too much before moving to Cursor, which meant more refactoring. The handoff point is roughly: I understand the shape of this. Not: this is mostly written.

---

The serious version

The site points to two resources: the Anti-Platform Positioning playbook and the SSV teardown with before/after examples.

The satire is the hook. The real argument is in those resources: sameness is structural, not a creative failure. Category pressure, approval chains, and timelines push companies toward the same words. The fix isn't a new adjective. It's specificity, proof, and being honest about who you're not for.

The implication for anyone in devtools or security GTM: if a solo builder with AI assistance can clone your category's messaging in a weekend, sameness isn't a brand problem. It's a moat problem.