Comparison
Compares explicit strategies, their benefits and costs, the closest alternative, and evidence that could reverse the working direction.
Editorial methodology
How a What If Paths decision lab is built
Each published page is designed as a small decision instrument, not a personality quiz. The lab asks for facts that can change the path, routes choices through a versioned graph, and produces an evidence brief whose format matches the decision itself.
Methodology version 4.0 · Last reviewed
One site, five output mechanisms
The topic determines the output. A job comparison should rank alternatives and show what would flip the direction; an emergency page should never end with a lifestyle score. Published labs currently use this distribution:
- runway
- 4 labs
- comparison
- 11 labs
- transition
- 2 labs
- consequence
- 2 labs
- emergency
- 5 labs
Transition
Separates hard gates from preferences. A failed legal, support, or readiness gate cannot be averaged away by enthusiasm elsewhere.
Runway
Uses entered cash, essential expenses, reliable income, and immediate shocks to calculate a transparent runway and likely first pressure point.
Consequence
Shows how exposure may unfold over ordered horizons, then identifies intervention windows rather than pretending there is one final outcome.
Emergency
Starts with danger triage, routes to the relevant situation, and returns a now/next/after action ladder plus official escalation boundaries.
How the model is constructed
- Define the decision boundary. Editors state who the page is for, who should not use it, what facts are outside scope, and when qualified or emergency help takes priority.
- Name real alternatives. Each lab defines topic-specific strategies rather than generic “safe,” “balanced,” or “bold” personalities.
- Choose context that can change routing. Context fields must alter a hard gate, branch, calculation, factor, or recommendation. Decorative questions are rejected.
- Author the graph. Standard labs contain a shared reasoning spine plus pressure and opportunity branches. Emergency labs use danger triage and five situation branches before coordination and reassessment.
- Write evidence-linked outputs. Every result exposes the entered context and selected choices that shaped it, plus three actions with a timing, rationale, and evidence check.
- Replay contrasting cases. Every published lab includes two editor-created composite cases that take meaningfully different routes and are recalculated through the same graph at build time.
Internal factor values help route a lab consistently, but public pages translate them into supported, exposed, or unclear. They are not probabilities, diagnoses, grades, forecasts, or measurements of a visitor.
Evidence, official sources, and editorial judgment
Sources support factual boundaries such as agency procedures, safety rules, benefit definitions, or legal/financial considerations. A source does not “validate” a personalized result. Each scenario lists the agency, jurisdiction, claim supported, and date checked so readers can see exactly what the reference does and does not establish.
Editors separately label judgment: trade-offs, sequencing, reversible tests, intervention windows, and stopping rules are editorial decision-support structures. They are written conservatively, reviewed against the source boundary, and never presented as a guarantee.
What earns a source
- Safety or emergency instructions
- Government, regulatory, or consular procedures
- Health, debt, employment, or immigration boundaries
- Claims that can change by jurisdiction or over time
What remains editorial
- How to sequence questions
- Which trade-offs to put side by side
- Which reversible test reduces uncertainty
- Which evidence should trigger reconsideration
Review, versioning, and quality gates
Decision graphs and saved progress are versioned. When a graph changes incompatibly, the browser does not silently reinterpret an old path as a new result. Published pages must pass structural checks for node references, branches, context rules, worked-case replay, mechanism-specific output, index policy, public-copy quality, and a production build.
Scenario pages carry a review date and revision history. A review checks broken or stale references, factual boundaries, graph validity, distinct case paths, mobile and keyboard use, and whether the result remains useful without implying certainty. Material changes update the graph version and public maintenance record.
Limits, privacy, and responsible use
Labs cannot observe a visitor's complete finances, health, contracts, relationships, location, legal status, or live danger. Results therefore remain working directions and evidence plans. Contact emergency services for immediate danger and use qualified professionals for financial, legal, medical, immigration, career, or mental-health decisions.
Interactive inputs and paths are stored locally in the browser so a visitor can continue on the same device. The lab does not require an account and does not intentionally send answer values to analytics. Aggregate events identify the route and mechanism—not the entered amounts or selected answer text. Resetting a lab or clearing site storage removes its local progress.