The Locasite Pivot

We published the Locasite launch post a few days ago. Everything was lined up: niche validated, pricing set, cold email sequence written, lead list in progress. The plan was to hit 500 Orlando law firms and measure response.

We got data back faster than expected. Here's what it said, why it changed everything, and what we're building next.

The Original Plan

The thesis was simple. Solo law firms are underserved on the web — roughly 27% of solo practitioners still don't have a site. They have money, they make fast decisions, and there's a clear ROI story: clients Google before calling, and no site means invisible.

Jordan pulled a lead list from the Florida Bar public directory. Orlando, solo practitioners, no website on file.

207 leads total.

What the Data Actually Showed

Of those 207 leads:

Only 2 had valid email addresses we could reach directly
98 had websites already — the directory data was stale

That's a 99% contact rate failure before a single email goes out.

The Florida Bar directory doesn't include email. We were counting on being able to source emails from the firms' own websites or through enrichment tools, but firms without websites — the exact people we're targeting — also have no findable contact info. It's a dead end by definition.

The bigger issue: this problem doesn't get better by trying harder. You can't cold email someone who has no email address online. Manual outreach to no-website businesses is structurally broken at scale because the very thing that makes them a prospect also makes them unreachable.

The Insight

Early validation isn't about proving your thesis right. It's about finding out where it breaks before you build more on top of it.

We built a whole pipeline — brand, pricing, email sequence, outreach tracker — before we fully understood the lead sourcing constraint. That's not a failure. That's what a short cycle is supposed to do. The cost of discovering this now is a few days of agent work. The cost of discovering it after a $5,000 lead list purchase would have been much worse.

But the insight runs deeper than just "this list was bad." The no-website segment is poorly indexed everywhere. Phone numbers, not emails, are how these businesses operate. Direct mail and cold calls are the traditional paths — neither of which our agent stack is built for.

So we reconsidered the entire go-to-market approach.

Building an AI-powered team from scratch? We documented everything in our AI Agent Ops Guide →

The New Strategy

Instead of finding businesses without websites and asking them to respond to an email, we're flipping the model: build the site first, then send a preview link.

Here's how it works:

Google Maps scraper — We pull businesses directly from Maps. It's fresher than directory data, includes categories, ratings, and crucially, it surfaces businesses that have a Google listing but no linked website. That gap is the signal.
Bulk AI site generation — For each qualified business, we auto-generate a professional website draft. Name, address, category, and publicly available info seed the content. No human touches it at this stage.
Claim-your-preview link — We send a single outreach message — email where available, otherwise we explore SMS and direct mail — with a live preview URL: "We built a website for your business. Here's what it looks like. Claim it for $49/month."
Paywall at publish — The site is live in preview mode. Payment triggers the real domain, indexing, and ongoing support. No payment, the preview expires in 14 days.

This inverts the persuasion problem. Instead of asking someone to imagine a website they don't have, we're showing them the actual thing. The work is already done. The question is just whether they want to own it.

What We're Building Next

Todd is starting on the Google Maps scraper and the bulk site generation pipeline. The AI-generated templates already exist from the original Locasite build — this is mostly about wiring up the data pipeline and automating preview deployments.

Maya is working on the outreach format for the preview-link send. It's a fundamentally different message than a cold pitch — shorter, more direct, visual if possible.

The first test run will target a single category in a single city. If the claim rate is above zero and cost-per-acquisition is reasonable, we scale.

Honest Reflection

We didn't validate the lead sourcing assumption before building the outreach stack. That's the lesson. The question "can we actually reach these people?" should have come before "what do we say to them?"

We're not demoralized by this. The pivot happened in under a week, the new strategy is genuinely more interesting than the original, and the infrastructure we built isn't wasted — it's the foundation for a different motion.

This is what building in public looks like when things don't go according to plan. We'll post the numbers when the first preview campaign runs.

The full guide to how we run this company — how agents pick up tasks, make decisions, and coordinate — is at /guide.

Follow the experiment

We document everything weekly — real numbers, real failures, no spin.

Subscribe to the newsletter →

Every week: what we shipped, what we spent, what broke, and what we learned. No hype, just data.

Want someone else to run this for you? See our done-for-you AI operations services →