← Back to blog
Architecture · Pricing

Shared cache: the unfair advantage making Xooriq 28× cheaper than Apollo

Apollo, Apify and ZoomInfo charge each customer for the same crawl. Xooriq charges once; all customers pay a $0.0006 cache hit. This post explains why this is unfair advantage (not just a feature) — and why Apollo can't copy without destroying their own pricing model.

1. "Unfair advantage" from Lean Canvas

In Running Lean (2010), Ash Maurya popularized unfair advantage — the last block of the Lean Canvas, usually the hardest. Literal definition: "something that cannot be easily copied or bought."

  • Tesla: Supercharger network + driving data (not buyable)
  • Stripe: developer-first DX + banking deals (years to replicate)
  • Notion: community + user-generated templates (network effect)

Shared cache for B2B data fits: structural architecture incumbents can't adopt without breaking their own business model.

2. Why Apollo can't copy

Apollo.io operates per-seat: $79/month/user (Custom). 5,000 paying customers × ~$500/month/company (multi-seat) = $150M ARR(source: GetLatka).

If Apollo adopted shared cache + $89/month flat:

  • 92% revenue loss — high-paying customers would drop to $99
  • Enterprise SLAs broken — Custom customers have "fresh data per seat" guarantees
  • Pricing power gone — market would price all B2B data below $50/seat

This is Christensen's Innovator's Dilemma: the structural incumbent cannot self-cannibalize. Xooriq can offer shared cache from day 1 because there's no legacy revenue to protect.

3. The cache hit math

Marginal cost per query

# Xooriq cache hit (PostgreSQL + Fernet decrypt)
Latency: 12ms (p50) / 38ms (p99)
CPU: 0.0001 cores
RAM: 4 KB
Cloud cost: $0.0006 per query

# Apify equivalent (compute unit)
Latency: 8-45 seconds (actor cold start)
CPU: 0.5-2.0 cores per actor
RAM: 512 MB - 2 GB
Cloud cost: $0.40 per CU

# Cost ratio
$0.40 / $0.0006 = 666× more expensive in Apify model

4. How Xooriq keeps cache fresh

Data typeTTLRefresh strategy
Legal name + CNPJ + address90 daysReceita Federal webhook when active
Estimated revenue + partners14 daysBNDES cron + Junta Comercial
Decision-maker + email + LinkedIn (public)14 daysFoxScraper institutional site
Job postings + tech stack24hDaily crawl Gupy, Vagas.com, Wappalyzer

5. Privacy: shared cache ≠ shared data

Common confusion: customers think "shared cache" exposes customer A's lead list to customer B. False.

# What IS shared (public, low risk)
- CNPJ 12.345.678/0001-90 → legal name, address, partners
- xooriq.com → tech stack: Next.js, Tailwind, PostgreSQL
- Job at xooriq.com → "Senior Backend Engineer · Remote"

# What is NOT shared (private, high risk)
- Your imported CSV list
- Your CRM annotations
- Your custom lead scoring
- Your outreach sequence
- Your sent emails

6. The moat over time

  • Day 1: cache covers 1M Brazilian CNPJs (top by search frequency)
  • Month 6: cache covers 5M CNPJs (regional B2B customers)
  • Year 1: cache covers 13.3M CNPJs (full active Receita Federal)
  • Year 2: cache includes MEI, ME, historic inactives

A new competitor would start from zero. Xooriq has a ~12-month cache lead.

7. Sources

Shared cache is our unfair advantage. Use it while Apollo sleeps.

$89/month · 13.3M companies · 666× cheaper than Apify · LGPD by design.

See full pricing