n8n Lead Generation Automation: Capture, Score and Notify in One Flow

The Webhook Trigger node generates a unique URL — something like https://your-n8n.com/webhook/lead-capture. Any form or service that can POST to a URL can feed this workflow. Configure each form tool's webhook setting to point here and you have a single entry point for your entire lead pipeline, regardless of how many forms or landing pages you run.

The first node after the webhook is a Set node that normalizes the incoming data. Different form tools name fields differently — Typeform uses form_response.answers, Webflow uses data.fields, a custom HTML form sends whatever you named the inputs. The Set node maps all of these to a consistent internal schema: lead.email, lead.name, lead.company, lead.message, lead.source_page, lead.utm_source. Every downstream node uses these standardized field names — which is why adding a new form source only requires adjusting the Set node mapping, not rebuilding the whole workflow.

Without a deduplication step, every time the same person fills out your form — revisiting after a few weeks, trying a different email, submitting from a second device — the workflow creates a new lead record. Your CRM fills with duplicates, your sales team contacts the same person twice, and your lead count metrics become meaningless. The fix is a single query before any write operation.

The Notion Query node (or Google Sheets lookup, or HubSpot Search node) filters your contacts database by the email field. If a result comes back, the lead already exists: the IF node branches to an "Update" path that appends a note to the existing record ("Resubmitted via /pricing page on Aug 11") and refreshes the Next Follow-up date. If no result, it continues to the new lead path. This two-branch structure means returning leads never create duplicates, but they do get their record updated — which matters because a second touchpoint is a strong buying signal worth noting.

Enrichment APIs cost money (or consume your free tier quota) per lookup. Running them against fake emails, mistyped addresses, and temporary inbox services wastes that quota and pollutes your CRM with uncontactable records. A validation step before enrichment filters the junk out first.

The HTTP Request node calls Hunter.io's Email Verifier endpoint: GET https://api.hunter.io/v2/email-verifier?email={email}&api_key={key}. The response includes a result field: deliverable (valid, continue), risky (might work, continue with caution), or undeliverable (invalid, stop). An IF node checks this field — undeliverable addresses branch to a dead-end path that logs the failed lead to a Google Sheet for review. Deliverable and risky addresses continue to enrichment.

Hunter.io's free plan includes 25 verifications per month. For higher volume, ZeroBounce ($0.008/verification) or NeverBounce ($0.008/verification) give accurate results at low cost — at 1,000 leads per month, that's $8, far less than the enrichment waste from skipping validation.

The enrichment step transforms a name and email into a full lead profile. The scoring formula in step 5 needs company size, industry, and seniority to work — without enrichment, you're scoring blind. The HTTP Request node calls Clearbit's free Enrichment API using the lead's email domain (everything after the @), which returns company name, industry, employee count, estimated revenue, headquarters country, and tech stack.

GET https://company.clearbit.com/v1/domains/find
Query param: domain = {{ $json.lead.email.split('@')[1] }}
Auth: Basic — username: YOUR_CLEARBIT_KEY, password: (empty)

The response gives you fields like metrics.employeesRange, category.industry, and geo.country. A Set node after the HTTP Request merges these into the lead object alongside the original form data. From this point on, $json.lead contains both what the person submitted and what the API returned — everything the scoring formula needs in one place.

For personal email domains (gmail.com, yahoo.com, outlook.com), Clearbit returns no company data. The Set node should default these fields to empty strings so the scoring formula doesn't throw errors — it simply assigns zero points for the firmographic signals it can't evaluate, which naturally pushes personal-email leads to a lower score.

This is where most guides go vague. "Add points for good signals" isn't a scoring system — it's a description of one. Below is a concrete scoring formula you can paste into n8n's Code node and adapt to your ICP. The formula covers firmographic signals (who they are) and behavioral signals (what they did), with negative points for disqualifying attributes.

const lead = $input.item.json.lead;
const enrichment = $input.item.json.enrichment || {};
let score = 0;

// --- Firmographic signals ---
const employees = enrichment.employeesRange || '';
if (['51-250','251-1000','1001-5000'].some(r => employees.includes(r.split('-')[0]))) score += 20;
else if (['5001-10000','10001+'].some(r => employees.includes(r.split('-')[0]))) score += 15;
else if (['1-10','11-50'].some(r => employees.includes(r.split('-')[0]))) score -= 10;

const targetIndustries = ['Software','SaaS','Technology','Marketing','E-Commerce','Finance'];
if (targetIndustries.some(i => (enrichment.industry || '').includes(i))) score += 20;

const country = enrichment.country || '';
if (['United States','Canada','United Kingdom','Australia'].includes(country)) score += 10;

// --- Title / persona signals ---
const title = (lead.title || lead.job_title || '').toLowerCase();
if (/director|vp |head of|chief|founder|owner|ceo|cto|coo/.test(title)) score += 20;
else if (/manager|lead|senior|principal/.test(title)) score += 10;
else if (/intern|student|junior|assistant|coordinator/.test(title)) score -= 20;

// --- Behavioral signals ---
const page = (lead.source_page || '').toLowerCase();
if (page.includes('/pricing') || page.includes('/demo')) score += 25;
else if (page.includes('/contact') || page.includes('/about')) score += 10;

const utm = (lead.utm_source || '').toLowerCase();
if (['google','linkedin','facebook','twitter'].includes(utm)) score += 10;

if ((lead.message || '').length > 80) score += 10;

// --- Negative signals ---
const personalDomains = ['gmail','yahoo','hotmail','outlook','icloud','proton','mail.ru'];
const emailDomain = (lead.email || '').split('@')[1] || '';
if (personalDomains.some(d => emailDomain.includes(d))) score -= 15;

const competitorDomains = ['competitor1.com','competitor2.io']; // add your own
if (competitorDomains.includes(emailDomain)) score -= 50;

// Clamp score to 0–100
score = Math.max(0, Math.min(100, score));

const tier = score >= 70 ? 'Hot' : score >= 40 ? 'Warm' : 'Cold';

return [{ json: { ...lead, enrichment, score, tier } }];

The formula is intentionally transparent — every point value is explicit, so when it gives the wrong result you know exactly why and can adjust. Adapt the targetIndustries, targetCountries, and competitorDomains arrays to match your ICP. After 2–3 weeks of running, compare your Hot leads against which ones actually converted — the scoring weights are your first calibration target.

Two chained IF nodes handle the routing. The first checks {{ $json.score >= 70 }} — true goes to the Hot branch. The false output feeds a second IF node that checks {{ $json.score >= 40 }} — true goes to Warm, false goes to Cold. The tier field already set by the Code node in step 5 serves as a readable label in all downstream messages and CRM records.

Hot branch: Three parallel nodes run simultaneously. (1) Notion or HubSpot creates a new contact with Status = "Hot / SQL", assigns the rep from your routing table, and sets Next Follow-up to today. (2) A Slack node sends a direct message to the assigned rep — not a channel message, a DM — with the lead's name, company, title, score, the page they came from, and their message verbatim. The rep receives this in Slack and can respond within minutes. (3) A task gets created in their task manager (Asana, Linear, or a Notion task database) with a due time of today at 3 PM. The combination of DM + task means no Hot lead gets missed even when the rep is in meetings.

Warm branch: Two parallel nodes. (1) Notion or HubSpot creates the contact with Status = "Warm / MQL" — no rep assignment yet, goes into a shared pipeline view that the sales team reviews weekly. (2) The lead gets added to a Warm nurture email sequence in Mailchimp, ConvertKit, or Brevo — 4–6 emails over 3 weeks covering use cases, customer stories, and a soft CTA. If they engage with the emails or visit your site again, a new webhook submission re-runs the scoring and potentially moves them to Hot.

Cold branch: One action only — log the lead to a Google Sheet with their email, score, and the timestamp. No CRM record (keeps your CRM clean), no sales contact (saves the team's time), but they do enter a long-form educational drip sequence (8–10 emails over 2 months). Some Cold leads convert after 6 months. The sheet gives you visibility without pollution.