noique/cross-border-ecommerce-skills/linkedin-prospecting

LinkedIn Prospecting

What this skill produces

A repeatable pipeline that turns one input — (product, target country, decision-maker role) — into a validated, contactable, message-ready list of LinkedIn decision-makers, complete with the email/phone the article-floor "google → linkedin" trick can't actually deliver:

1. A search plan of 8–15 Google + Bing + Wayback dorks adapted to the country's primary LinkedIn profile language and to the role's local title
2. A profile list with LinkedIn URL, name, title, company, location, mutual-connection count, post-activity tier
3. An enrichment file with email + phone (where legally extractable) via Apollo/Snov/Hunter/Lusha/Wiza
4. A 4-touch outreach sequence — Connection Request → Accepted DM → Follow-up → Channel-switch — with templates in EN / ES / PT / FR / DE / AR / JA, plus a reply-handling matrix
5. A compliance pass — LinkedIn ToS limits + GDPR / CASL / CCPA flags per country

The end state: a junior salesperson can pick up the list and start sending without rewriting messages, and the LinkedIn account doesn't get rate-limited or restricted in week one.

When to use this skill

• Foreign-trade seller wants to skip LinkedIn Sales Navigator ($99/mo) and still find the same people
• Founder needs 50–200 cold conversations with overseas decision-makers to validate a market
• Agency is building an outbound list of procurement/buying contacts for a manufacturer client
• A previous WhatsApp campaign (google-whatsapp-prospecting) is exhausted — switch to LinkedIn channel for the same buyer set

Do NOT use for:

• B2C influencer outreach — see tools/backlink-kol-extractor and brand-strategy/influencer-marketing
• Recruiting / hiring — different ToS rules apply, different role keywords
• Markets with low LinkedIn penetration — Russia (LI is blocked), China mainland (LI shut down 2021), Iran. Switch to local platforms.

Required inputs

Refuse to start with vague inputs.

Optional: target company-size band, industry vertical, language requirement (EN-only vs OK with PT/ES native).

The four-stage pipeline

Each stage has a dedicated reference. Read it before executing the stage.

Stage 1 — Search

Build a query plan from references/search-formulas.md. The article's single formula

site:linkedin.com/in/ "<product>" "<country>" "<role>"

is the floor — it now misses 30–60% of profiles because of LinkedIn's 2024 Auth Wall and Google's resulting deindexing. Real campaigns add:

• Bing + Yandex mirrors (often have less LI deindexing than Google in 2025–26)
• Localized role keywords from references/role-keywords.md — "Comprador" for Brazil, "Acheteur" for France, "Einkaufsleiter" for Germany — most countries' LI profiles use the local title
• Wayback Machine fallback for profiles Google has dropped: https://web.archive.org/web/*/linkedin.com/in/*<role-keyword>*
• Company-page → people chain: find target companies first via site:linkedin.com/company/, then dork their employees
• Post-activity targeting via linkedin.com/posts/ to find people who posted about the product recently
• Exclusion of consultants/freelancers/recruiters who pollute results: -"open to work" -"consultant" -"recruiter" -"hiring"

Stage 2 — Enrich (find email + phone)

Article A vaguely promises "tools to extract emails" — references/enrichment-tools.md names them with cost, accuracy, and legal context. The realistic stack:

• Free tier: Hunter.io (50/mo), Snov.io (50/mo), Apollo.io (60/mo) — start here for under-50 lists
• Paid bulk: Apollo ($49/mo, 12k credits), Lusha ($29 starter, mobile-first), Wiza ($30, LinkedIn-export specialist)
• Manual fallback: pattern-guess email from company domain (firstname.lastname@company.com) → verify with NeverBounce / ZeroBounce

Tools cost money. The article elides this — set client expectations.

Stage 3 — Outreach (4 touches)

Use templates/outreach-playbook.md. The article gives 7 DM templates, all of one archetype (pain → solution → low-friction CTA). The playbook adds:

• 5 more archetypes the article missed: mutual-connection intro, replied-to-post warm, recent-news/award reference, reverse-pitch ("we want distributors"), voice-note opener (highest 2026 reply rate)
• Connection Request template — completely separate from DM, hard 300-char limit, different psychology
• Per-language native templates — not Google Translate
• Reply-handling matrix — what to send when they ask price / send catalog / not interested / ghost

Touch cadence:

1. T0: Connection Request (≤300 chars)
2. T1 (24–48h after acceptance): Opening DM using one of 12 archetypes
3. T2 (5d after T1, no reply): Follow-up
4. T3 (10d after T2, no reply): Channel switch — pivot to email or WhatsApp using enrichment from Stage 2

Stop after T3. Do not chase further on LinkedIn — your account gets flagged.

Stage 4 — Compliance + LinkedIn account hygiene

Read references/linkedin-limits.md before any campaign >25 connection requests/day. Hard limits in 2026:

• Connection requests: ~80–100/week (down from 200/week pre-2024). Exceed → "You're approaching the weekly limit" warning, then 1-week freeze.
• DM volume: ~100/day for 1st-degree connections, 0/day for 2nd+ unless you pay InMail credits
• Automation tools (Phantombuster, Linked Helper, Dux-Soup, Octopus, Expandi): all violate LinkedIn ToS § 8.2. Account ban is the standard outcome — even "throttled" tools eventually get caught.
• Country regulation: same GDPR / CASL / CCPA / UWG §7 as google-whatsapp-prospecting — LinkedIn DM is "commercial electronic message" under most regimes

Recommended account hygiene:

• One personal LinkedIn per active outbound campaign
• Warm the account for 2 weeks (post 1×/week, comment 3×/week, accept inbound) before high-volume outbound
• Never send the same message text >5 times — LI similarity-detects across accounts

Workflow (end-to-end)

1. Confirm the three required inputs. Refuse to proceed if vague.
2. Open references/role-keywords.md, get the localized role list for the target country.
3. Open references/search-formulas.md, build 8–15 queries combining: localized role + product + country + Google/Bing/Wayback + exclusions.
4. Execute searches manually or via SerpAPI batch (reuse templates/serpapi-batch.py from google-whatsapp-prospecting/templates/). Dedupe LinkedIn URLs.
5. For each LI URL, capture: name, current title, company, city, mutual connections (if visible without login), post-activity tier (last 30d).
6. Run the URL list through Apollo / Snov / Hunter — see references/enrichment-tools.md for picking the right tool by region. Get email + (where available) mobile.
7. Score each lead (templates/lead-tracker.csv's priority column): mutual connections (×3), post-activity (×2), title match (×2), company size match (×2).
8. Write Connection Requests for top 80 leads/week using templates/outreach-playbook.md.
9. After acceptance (T1): pick the right DM archetype based on what you saw on their profile. Personalize the merge fields. Send only inside their local 09:00–17:00.
10. Track in CSV. Run T2 follow-up at 5d. Run T3 channel-switch at 10d (use email from Stage 2, or WhatsApp number if surfaced).
11. After T3, archive. Do not chase further on LinkedIn.

Common failure modes

• Using English role titles in non-English markets: searching "Purchasing Manager" "Brazil" returns ~25% of the BR procurement population that lists their title in English. Always include the local term (Portuguese: "Gerente de Compras" / "Comprador").
• Hunting CEOs of large enterprises: in companies with 1000+ employees, the CEO never signs purchase orders. Filter to small/mid companies (10–500 employees) when targeting CEO/Owner; for enterprise, target Procurement/Sourcing director.
• Connection request opens with "Hi, hope you're doing well": 300 chars are precious. Cut all pleasantries, lead with the specific reason in 1 sentence.
• Sending 7 identical templates from the article: LinkedIn's similarity detection flags pattern-blast accounts. Use the templates as scaffolds, rewrite the verb and the example each send.
• "Free gift / send link" hook in T1: triggers LI's auto-spam classifier. Defer the link to T2 after they reply.
• Skipping enrichment: 30% of leads will ghost on LI but reply on email. Without an email column you can't do the T3 channel-switch.
• Running 200+ connection requests/day via Phantombuster: account gets restricted in 7–14 days. Worth it only if you're OK losing the account.
• Sending DM at 3am buyer-local time: same time-zone issue as WhatsApp prospecting; cross-reference google-whatsapp-prospecting/references/country-targeting.md for send windows.

Source-article gap analysis (2026-04-26)

Source articles built from:

• "开发信没邮箱？一键复制这个Google公式，精准反查LinkedIn海外买家CEO！" (慧进AI工具说)
• "建议马上收藏！外贸老鸟都在悄悄用的7个LinkedIn高回复私信模板" (慧进AI工具说)

See also

• references/search-formulas.md — Google + Bing + Wayback + company-chain dorks for LI
• references/role-keywords.md — 50+ decision-maker titles in 8 languages, by company-size band
• references/enrichment-tools.md — Apollo/Snov/Hunter/Lusha/Wiza comparison + free fallback patterns
• references/linkedin-limits.md — quota table + ToS § 8.2 (automation ban) + account warm-up protocol
• templates/outreach-playbook.md — Connection Request + 12 DM templates + T2/T3 sequence + reply matrix
• templates/lead-tracker.csv — CRM with priority scoring
• Sister skill: ../google-whatsapp-prospecting/ — same 4-stage shape but channel = WhatsApp; share the SerpAPI script