Trilha Certa

Methodology Policy — Personalized Rates vs Public Data

Why this policy exists

For products like personal loans, many institutions only reveal final rates/CET after user-specific analysis (account relationship, score, profile). This creates a hard boundary between:

  1. public, reproducible information
  2. personalized, user-dependent outputs

Core rule

Never present personalized outputs as market-wide objective rankings.

Two-layer model

Layer A — Public Baseline (publishable as objective)

Use only public, reproducible sources:

  • official product pages
  • official contracts/terms/PDFs
  • regulatory references (e.g., BCB)

Allowed outputs:

  • ranges
  • examples explicitly marked as examples
  • contractual rules (tarifa/IOF policy, prazo limits)

Layer B — Personalized Simulation (user-specific)

Use simulator/app outputs that depend on user profile.

Allowed outputs:

  • user-specific comparison view
  • explicit profile context
  • no claims of universal ranking

Publication rules

Can publish definitive cross-bank ranking only if:

  • critical numeric fields are resolved from comparable, non-profile-dependent sources OR
  • personalized outputs are collected under an explicitly equivalent profile framework (same profile assumptions and transparent limitations)

Must not publish definitive ranking when:

  • core fields depend on account-only personalization and cannot be normalized fairly.

In that case publish:

  • comparison framework
  • evidence registry
  • methodology explanation
  • “ranking pending profile-specific inputs” status

Required disclosure block (for relevant pages)

"As condições de empréstimo pessoal podem variar por perfil, relacionamento e análise de crédito. Quando os dados disponíveis forem personalizados, o Trilha Certa apresenta esses valores como simulação específica, não como ranking universal de mercado."

Process impact

  • Task 03 outcome can be CONDITIONAL PASS with publication mode set to:

    • framework-only (no definitive ranking)
    • personalized-comparison (profile-scoped)
    • market-ranking (only when strict comparability gate is met)

Stop rule (accepted by product policy)

When key rates are personalized and cannot be normalized fairly, the process should stop at framework-only publication mode. No forced direct cross-bank ranking is required in this case.