BlipLeads

The Challenge

Extracting data from Google is the easy part. The real challenge is turning it into a B2B sales pipeline that supports multiple clients concurrently without hitting API limits, mixing client accounts, or stalling on slow queries:

• / Processing long-running searches (Google Places + website probing) without blocking HTTP requests, using asynchronous background jobs.
• / Isolating data per tenant (campaigns, leads, metrics) and by user role (owner/admin/operator/viewer).
• / Preventing duplicate leads per tenant even with multiple campaigns running for the same niche and region.
• / Automatically detecting 'dead' websites listed on Google, turning a false positive into a sales opportunity.

The Solution

The workflow is campaign → job → lead. Creating a campaign publishes a job to Upstash QStash, triggering the search asynchronously — without freezing the page. The routine normalizes phone numbers, verifies whether listed websites are actually online, calculates priority scores, and persists everything with account-level unique indices. Two clients running campaigns for the same niche will never duplicate leads.

export function calculatePriorityScore(input: ScoringInput): number {
  let score = 0;

  if (!input.websiteUri) score += 30;

  if (input.rating != null) {
    score += (input.rating / 5) * 20;
  }

  if (input.reviewCount != null && input.reviewCount > 0) {
    score += Math.min(25, Math.log10(input.reviewCount + 1) * 15);
  }

  if (input.businessStatus === 'OPERATIONAL') {
    score += 5;
  }

  return Math.round(score * 100) / 100;
}