Hank finds local businesses with failing websites, builds their new site before they ever ask, and signs them up — no sales team, no cold calls, almost no human hours. We test it together with a small month-one experiment; if the numbers show up, we scale it and sell it.
Drive twenty minutes past Lake St. Louis and count the businesses with no website, a dead Facebook page, or a site built in 2011 that takes eight seconds to load on a phone. These are good businesses — plumbers, dentists, HVAC crews, restaurants — losing customers every day to whoever shows up first on a phone screen.
And the ground just shifted under them a second time. People increasingly ask ChatGPT, Perplexity, and Google's AI for recommendations instead of clicking links. Those AI tools read structured, machine-readable websites. A business with a bad site was losing the Google game; a business with a bad site is completely invisible in the AI game. Almost nobody is selling them the fix — least of all at a price a small shop says yes to.
The brand is already real: Hank — the friendly neighbor who happens to be great with computers. "I made you some things." Warm, plain-spoken, zero jargon — built for a business owner who feels left behind by tech. hankbuilds.com is registered, the identity system is done (that's him up top), and every asset below is genuinely generated by the pipeline, not concept art.
Figures are directional planning estimates, not audited market research — the pilot exists to replace assumptions with data.
We start in the St. Louis exurbs — where web presence is weakest and nobody's competing for these customers — and expand metro by metro using the exact same playbook. The pipeline doesn't care what city it's pointed at. Every metro is a fresh territory of thousands of qualifying businesses, and the qualifying criteria are objective and automated (Google's own speed scores, missing sites, no security certificate — not opinions).
This is not a "win a huge market" bet. It's a "nobody is serving these people at this price point, and Hank can serve them nearly for free" bet. Our cost to find, evaluate, build a site for, film a personalized pitch for, and mail a postcard to one business: about $1.50.
Then the real money: upsells into the accounts we've opened. A monthly AI-visibility report ($19–29), content refresh and review monitoring (bundled ~$79–99 tier), and the flagship — an AI phone receptionist that recovers the calls these businesses miss every day ($100–300+/month, a proven category). The $39 site is the foot in the door; the ladder is the business.
Not concept art — generated end-to-end by the machine for real qualifying businesses (Ellisville Heating & Air, Henderson Heating & Air, Lake St. Louis Heating and Cooling, Troy Heating & Cooling, George Littleton's Plumbing), plus category demos:
The builder doesn't pour every business into one mold — you can see it in the gallery above. Design adapts to business type: a plumber's site leads with a giant tap-to-call and "24/7 emergency" above the fold, because their customer is standing in a flooded basement; a restaurant leads with menu, hours, and photos; a dental office gets a calm, clinical layout with booking front and center. Colors, tone, layout, and calls-to-action all key off the business's category — pulled from the same Google data that found them.
And it learns from the winners. The same grading engine that catches bad sites also studies the highest-scoring sites in each category — what the 90+ sites do with structure, speed, content, and CTA placement — and those patterns are folded back into the templates. Then our own portfolio becomes the second teacher: every preview site reports what visitors actually do (scroll, tap-to-call, buy), so each monthly batch ships templates that convert better than the last. That compounding feedback loop is a real asset at exit time — an acquirer isn't just buying customers, they're buying a design system trained on what makes local businesses convert.
The gate code rotates every 24 hours — I'll send you today's. You'll experience exactly what a business owner experiences.
Two clocks run at once. The batch clock: every 30 days, a fresh batch of ~500 scored businesses enters the pipeline — sites built, videos filmed, first postcards mailed — triggered on a schedule, capped by hard spend limits. The per-business clock: from the moment their postcard drops, each business walks a behavior-driven mail sequence inside their 60-day preview window. Every scan, conversation, and checkout is an event the engine sees — so nobody gets a "reminder" for something they already did, and buyers exit the sequence the second they subscribe.
Print-ready proofs from the actual postcard generator — each one is built per-business with their name, their preview URL, and their unique code.
Under the hood it's a simple state machine per business — mailed → scanned → talked → subscribed / expired — with the gate, Hank, and checkout each reporting events, and the mail API triggered automatically when a business sits in a state past its timer. Every send checks the monthly spend cap first; the cadence can never outrun the budget. Cost honesty: worst case a business gets three cards, which moves estimated acquisition cost from ~$55 toward ~$75–85 if conversion didn't improve at all — but multi-touch sequences exist precisely because they lift response more than they lift cost. The pilot measures exactly that.
Working assumptions (to be validated by the pilot, not taken on faith): about 3 in 100 mailed businesses become customers — 30% scan the postcard, half of those talk to Hank, a fifth of those buy. Even if we're half wrong, acquisition cost doubles to ~$110 and still pays back inside three months. The funnel is cheap enough to be wrong about and still win.
| Stage | Customers | Revenue (ARR) | Gate to advance |
|---|---|---|---|
| Pilot (mo 1–3) | ~15–45 | ~$7–21K | Real conversion & churn data from 500 businesses mailed in month one and each month after. Why 500, not 100: a 100-card test can't statistically tell a 1% funnel from a 6% one — 500 can. ~$1,100–1,300/mo to run. |
| Validation (mo 4–6) | ~60–90 | ~$40K | Four-part trigger: conversion ≥2%, cost-per-customer ≤$110, churn ≤7%/mo, AI handles 90%+ of conversations. All four hold → expand immediately. |
| Multi-metro (yr 2) | ~1,000 | ~$640K | The playbook replicates in 2–3 new metros at similar cost. Part-time VA runs the exception queue. |
| Profit target (yr 3–4) | ~2,400 | ~$1.5M | $1M/yr net profit run rate. One real hire absorbing operations. |
| Exit-ready (yr 4–5) | ~5,000 | ~$3M | $2M+ profit, documented AI-run operations, clean retention metrics. |
Businesses like this sell for roughly 3–5× profit. $1M/year of profit fetches $3–5M; the $10M exit needs roughly $2M+ in annual profit — about 5,000 customers — or premium metrics that justify a higher multiple. Here's our edge on the multiple: buyers pay premiums for businesses that don't need the founder. A company with documented, provable "AI runs everything, humans touch exceptions only" operations — plus a beloved consumer-facing brand like Hank — is the most transferable asset you can sell. Our lifestyle constraint (nearly hands-off) and our exit strategy (premium multiple) are the same strategy.
Default path: bootstrap. 100% of profits reinvested until 5,000 customers. No investors, no dilution, we own the whole exit. The numbers above don't require outside capital — the machine funds itself after validation.
Where VC would make sense (and only here): if the pilot numbers come back meaningfully better than assumed — say conversion at 5%+ with low churn — then speed becomes the play: raising $1–2M to blitz 20 metros before anyone copies the playbook could turn a $10M exit into a $30–50M one. That's a good problem, we'd decide it together with real data, and it's the only scenario where giving up ownership beats keeping it. VC also makes sense if a strategic acquirer (a marketing agency roll-up, a local-services platform) shows early interest — capital accelerates toward a known buyer.
Where selling early makes sense: if we get to ~1,000 customers and an acquirer offers 4–5× ARR (~$2.5–3M) — that's a life-changing return on a $2K experiment and three years of light-touch work, and saying yes is allowed. We don't have to be heroes about the 10.
Structure: deliberately decided later. The trial is just a trial — we don't paper an equity split over a $2K experiment. If the pilot numbers hit the triggers and we both want in, then we sit down and structure it properly in writing (equity vs. capital-plus-share, vesting, the works — with an hour of a real attorney's time, because friends who write things down stay friends). You get full visibility into every number the whole way: same dashboards I see.
You're a friend before you're a partner, so here's the whole truth:
Hank Builds · hankbuilds.com · St. Charles County, MO