Lead sources — the NO-SITE lane

The weak-site email lane can't reach businesses with no website — because they usually have no findable email either (Oliver's "egg or hen paradox"). This lane finds them off-web and gets a phone number, then makes them reach out to us — a QR-to-claim, a public comment, a scan. That single design choice keeps consent inherent and every market 🟢.

~£0.59cost / verified lead (turk)
~£36cost / acquired customer
£300–500rank-and-rent anchor / mo
Postalcompliant everywhere 🟢

The organising principle — "make them reach out to us"

Consent is inherent

However we source a number, we use it for one postal touch that ends in the prospect self-initiating — scanning the <slug>.ozdreamsites.de QR, commenting on a public post, or phoning us. No cold call, no cold SMS.

That sidesteps PECR (UK) and UWG §7 (DE) entirely — the number funds a letter, never a call.

The binding constraint isn't cost

Sourcing a no-site lead is cheap (~£0.59). The real metric is claim-rate — how many scan the QR and convert. So we prove the claim-flow with free channels before paying to source at scale.

The five channels

① Meta / Facebook — local groups

No-site trades live in town buy/sell, "recommendations", trades & mums groups. A post with a flyer image + number and no URL is the qualifying signal.

Play: public "comment for a free preview website" post → they comment = they initiated. Flyer images → OCR → postal only.

🟢 inbound posts & ads · 🔴 no cold DMs, no logged-in scraping (Meta ToS + GDPR).

② Newspapers & classifieds

Pre-qualified no-site lists with numbers already printed: Gumtree, Marketplace, Kleinanzeigen.de, Nextdoor, local-paper back pages, parish mags, DE Wochenblatt.

Play: human-browse → parse numbers → postal A-pile + QR.

🟢 → postal (UK) · 🟠 DE postal-only — a listed number is not phone-consent (Saarland OVG).

③ Community boards — "mechanical turk" ⭐

The richest data is off-web entirely: newsagent windows, supermarket boards, "white van man" cards. Pay a local "teen with a bike" (v1: adults 18+) per photo / per verified lead; AI OCRs name + number → CRM.

v1: WhatsApp intake → n8n → Google Vision + GPT-4o → dedupe → CRM → Stannp postcard + QR.

🟢 → postal · unique, un-competed data — the scalable moat.

④ Sign-writers / vehicle livery

Whoever painted the number on every local van holds a de-facto contact list of our exact target.

Play: approach 1–2 as referral partners (fee / rev-share; their customer opts in). Automated twin = Street View / shopfront OCR (v2).

🟢 referral · 🔴 no raw list-buying without a full data-sharing agreement.

⑤ Compliance spine

Sole traders = individual subscribers under PECR → cold e-mail/SMS needs consent. DE B2B calls need presumed consent; a listed number doesn't count. A phone number is personal data (GDPR) — basis = legitimate interest + LIA + minimisation + erasure.

Everything funnels to postal + QR-to-claim, which is 🟢 in both markets on Oliver's compliance.ts.

Mechanical-turk unit economics (worked example)

StepAssumptionCost
Bike round40 photos in ~2 hrs40 × £0.25 base = £10.00
Extraction~5 usable leads/photo → 200 raw → 120 unique newVision £0.05 + LLM £0.15 ≈ £1
Lead bonus120 new verified leads120 × £0.50 = £60.00
Postal touch1 Stannp A-pile postcard + QR each120 × £0.50 = £60.00
Total120 targeted no-site, QR-to-claim prospects≈ £131
Cost / verified lead£71 sourcing ÷ 120≈ £0.59
Convert @ 3% claim~3.6 paying customersCAC ≈ £36
~£36cost / customer
£360–4801st-yr LTV @ £30–40/mo
~10×first-year return
£35/hrwe can pay the worker & still profit

Even against the £300–500/mo rank-and-rent anchor, acquiring a subscriber for ~£36 prints money — and we can pay bike-round workers generously. The economics work comfortably; claim-rate is the constraint, not sourcing cost.

Compliance — one table to rule them all

ActionUK (PECR + GDPR)DE / DACH (UWG §7)Verdict
Postal A-pile + QR-to-claimNot covered by PECR; legit interestPost 🟢; UWG is e-/phone ads🟢 primary
Inbound FB post → they comment/DM usThey initiate = consentThey initiate = consent🟢
Boosted / lead-form adsOpt-in enquiryOpt-in enquiry🟢
Sourced number → postal touchLegit interest (published no.)Postal only; listed ≠ consent🟠 postal-only
Sign-writer referral (customer opts in)Consent via partnerConsent via partner🟢
Cold call to sourced numberTPS/CTPS + consentPresumed consent; listed ≠ consent🔴 avoid
Cold SMS / email to sourced numberSole traders = individuals → consentPrior express consent🔴 avoid
Logged-in scraping of Meta / GoogleToS breach + GDPRToS breach + GDPR🔴 avoid
Paying under-16s to photographCouncil permit, hour caps, safeguardingJugendarbeitsschutz🔴 v1: 18+ only

Cheapest-to-test-first — prioritised order

1Inbound Facebook post

£0, 🟢, testable today in one town. "Comment for a free preview website." Proves demand + the claim-flow at zero sourcing cost. Start here.

2Classifieds → postal

Gumtree / Kleinanzeigen / Nextdoor / paper back-pages. Near-£0, numbers already published, 🟢-postal. Fills the postcard pipeline fast.

3Mechanical turk (WhatsApp v1)

~£131 pilot, 2 adult workers, one town. 🟢-postal. Unique un-competed data — the moat. Turn on once the claim-flow is proven.

4Sign-writer referral

£0 upfront, 🟢, but relationship-paced. Run in parallel as a warm channel.

5Boosted / lead-form ads

£5–10/day, 🟢, optional amplifier once the messaging is proven.

6Street View / shopfront OCR

Build cost + GDPR care, 🟠. v2 experiment — lowest priority; official APIs / manual capture, no bots.

Sequencing: prove the claim-flow free (① + ②) → turn on the paid sourcing moat (③) and warm partner (④) → amplify with ads (⑤) → automate with OCR at scale (⑥).

Klaudius · Lead sources — the NO-SITE lane · full strategy in outreach/lead-sourcing-strategy.md · research draft v1