checkads.
Transparency & Scientific Rigor

Data Methodology & Provenance Assurance

How checkads aggregates, verifies, normalizes, and cryptographically preserves growth intelligence without compromising business privacy.

1. The 4 Verification Statuses

Every metric published on checkads strictly belongs to one of four mutually exclusive statuses. We never blend or mislabel unverified figures.

VERIFIED

Directly authenticated via authorized OAuth API feeds (Meta Ads, Shopify, Stripe). Raw JSON responses are hashed with SHA-256 for cryptographic audit trails. Account owners cannot manually edit verified figures.

ESTIMATED

Calculated deterministically using verified inputs where partial attribution exists (e.g. Blended CAC = Verified Spend / Verified New Customers, or Platform ROAS).

SELF-REPORTED

Data provided voluntarily by company representatives without automated API verification (e.g., historical founding year or team headcount).

UNAVAILABLE

Data source disconnected, integration revoked, or metric sample size insufficient. We never fabricate placeholder numbers.

2. Core Metric Formulas

Blended Marketing Efficiency Ratio (MER):
MER = Total Net Revenue / Total Advertising Spend
Guarded against zero ad spend. Distinguishes blended commercial efficiency from platform-attributed ROAS.
Platform Attributed ROAS:
ROAS = Conversion Value Reported / Ad Spend Reported
Restricted strictly to authenticated Meta Ads pixel/CAPI attribution windows.
Customer Acquisition Cost (CAC):
CAC = Total Marketing Acquisition Spend / Verified New Customers Acquired
Returns null if customer acquisition data is unavailable or zero.
Period-over-Period Growth:
Growth = ((Current Period Revenue - Previous Period Revenue) / Previous Period Revenue) × 100%
Handled with zero-denominator safeguards and negative revenue bounds.

3. Benchmark Anonymity & $N \ge 10$ Privacy Moat

Aggregated industry benchmarks serve as an industry public good while guaranteeing absolute company confidentiality:

  • Strict Sample Threshold: No benchmark cohort displays aggregate numbers unless at least 10 independent verified companies are active ($N \ge 20$ for sensitive sub-cohorts).
  • Zero Inferential Leakage: Quartiles are computed using Type-7 interpolation with outlier Winsorization, preventing single-company influence on benchmark percentiles.
  • Frozen Snapshots: Snapshots are frozen monthly. Dynamic updates cannot be diffed to reconstruct individual business changes.