I built an n8n workflow that finds local businesses with high Google visibility but terrible websites. The bigger the mismatch, the hotter the lead. It's like arbitrage for web consultants.
The Problem with Traditional Lead Gen
Most B2B prospecting tools scrape business directories, LinkedIn, or email databases. You get polished companies with polished websites who already have agencies on retainer.
The real opportunity? Businesses thriving on Google Maps with 2005-era websites.
A plumber ranking #3 in Helsinki with a broken contact form isn't hard to find. They're just invisible to traditional lead gen tools.
The Mismatch Hypothesis
High search visibility + Low web quality = High purchase intent
These businesses already have customers. They just don't know how much revenue they're leaving on the table with poor online presence.
They're not "prospects." They're pre-qualified buyers.
How the Workflow Works
The n8n template executes a four-stage pipeline:
Stage 1: Google Maps Search
- Serper API searches Google Maps for target niche + location
- Returns businesses ranked by Google's local algorithm
- Captures ranking position as "visibility score"
Stage 2: Website Scraping
- Firecrawl scrapes each business website
- Extracts full page content, structure, metadata
- Handles JavaScript-heavy sites, SPAs, dynamic content
Stage 3: AI Analysis
- Groq (llama-3.1-70b-versatile) analyzes website quality
- Scores against 8 criteria: mobile responsiveness, load speed, contact clarity, trust signals, content depth, visual design, SEO basics, conversion path
- Each criterion scored 0-10
Stage 4: Mismatch Calculation
- Compares Google visibility vs website quality
- Formula:
Mismatch Score = (Google Rank / 20) * 10 - (Average Quality Score) - Higher score = bigger opportunity gap
- Exports to CSV with business name, website, phone, scores, AI reasoning
Why This Works
Targeting logic:
- Business ranks well → Has customers, generates revenue
- Website scores poorly → Knows they need help, hasn't solved it yet
- High mismatch → Maximum pain point, minimum competition from other agencies
Real example from Helsinki:
- HVAC company, Google Maps position #2
- Website: Flash-based, broken on mobile, no online booking
- Mismatch score: 8.7/10
- They took my call. They signed a contract.
Technical Architecture
// Simplified node flow
Google Maps Search (Serper)
↓
Loop Over Results
↓
Firecrawl Scrape
↓
Groq AI Analysis
↓
Mismatch Calculation
↓
CSV Export
Key n8n patterns:
- Error handling on scrape failures (some sites block bots)
- Rate limiting between Firecrawl calls
- Conditional logic to skip businesses without websites
- Dynamic AI prompting based on business type
API Costs (Per 100 Leads)
- Serper: $0.50 (100 searches)
- Firecrawl: $2.00 (100 scrapes at $0.02 each)
- Groq: $0.00 (free tier handles this easily)
Total: ~$2.50 per 100 qualified leads
Compare that to $50-100 per qualified lead from traditional B2B databases.
Use Cases Beyond Web Design
This workflow adapts to any local service vertical:
- SEO agencies: Find ranked businesses with poor technical SEO
- Booking platforms: Target service businesses without online scheduling
- E-commerce consultants: Identify retailers with outdated payment flows
- Marketing agencies: Score businesses with high traffic, low conversion
Just adjust the AI scoring criteria in the Groq prompt.
Results from 30 Days
I've been running this weekly for Helsinki B2B services:
- Leads generated: 340
- High-mismatch (score >7): 89 (26%)
- Outreach response rate: 34% (vs 8% industry average)
- Booked calls: 19
- Closed deals: 4
The mismatch score is predictive. Businesses scoring 8+ respond at 3x the rate of random cold outreach.
Get the Template
The full workflow is available on the n8n Creator Hub:
Find High-Mismatch Local Business Leads with Firecrawl and Groq
It's free to download and customize. You'll need:
- n8n instance (self-hosted or cloud)
- Serper API key (free tier works)
- Firecrawl API key
- Groq API key (free)
Customization Ideas
For agencies:
- Add webhook trigger for weekly automated runs
- Connect to CRM (HubSpot, Pipedrive) for auto-import
- Add email enrichment node to find decision-maker contacts
For consultants:
- Filter by business age (newer = less technical debt)
- Add review sentiment analysis (high reviews + bad site = urgent need)
- Score competitor websites for competitive positioning
For developers:
- Replace Groq with Claude/GPT for more nuanced analysis
- Add screenshot capture for visual site audits
- Build PDF report generator for prospecting materials
Why I Built This
I'm an n8n Ambassador and workflow automation consultant. Most lead gen workflows I see are just scrapers with extra steps.
This one actually thinks.
The mismatch score isn't arbitrary. It's measuring a real business problem: the gap between a company's market position and their digital infrastructure.
That gap is money on the table. For them, and for you.
What's Next
I'm testing two additions:
- Competitor analysis node: Scrape top 5 competitors in same niche, compare feature sets
- Personalized outreach generator: AI writes custom email based on specific site weaknesses
If you're using this workflow, I'd love to hear what verticals you're targeting and what mismatch scores you're seeing.
Drop a comment or find me on LinkedIn.
Mychel Garzon
Workflow Automation Consultant
Helsinki, Finland
