VARIARE LAB · TRUST & TRANSPARENCY

Methodology & standards

Math in the open. Models begin with explicit assumptions and a defined question. Each model documents its own formulas, units, supported inputs, limitations, and examples alongside the result.

All four public models use deterministic calculations: the same valid inputs produce the same unrounded results. Their outputs can be reproduced from the published formulas. Automated tests cover known examples, invalid inputs, boundaries, and relevant edge cases. These checks are technical validation, not evidence of commercial usefulness or guaranteed real-world outcomes.

The three Explorers compare one changed assumption while keeping the others constant. Hiring compares the same company without new hires against a hiring decision defined by count, loaded cost and start month. Company assumptions remain the same in both paths. Labels identify what changed and how the result differs; comparison does not establish what would happen in the real world.

Models have different scopes. Inflation & Purchasing Power holds a nominal balance fixed under an assumed constant inflation rate. Real Return models constant nominal investment returns and inflation. Break-Even models monthly price, fixed costs, and variable cost per unit. Hiring projects monthly company cash and runway; its direct hiring effect is additional payroll, with no automatic revenue or productivity uplift. Read each model’s methodology for the exact formulas and exclusions.

Calculations retain internal numerical precision; rounding is applied for display and does not feed back into the model. A displayed rounded result may differ from a calculation using already-rounded intermediate values. Whole-unit break-even results round upward. Hiring reports approximate runway while retaining internal precision; its requirements search a bounded model on a cent or whole-month grid, as documented on the page.

Hiring’s What matters most? tests change one input around the hiring scenario and compare minimum cash over the selected horizon. These bounded local tests are not probabilities or universal measures of importance, and their effects cannot be added together. What must be true? finds additional starting monthly revenue OR a later hiring month needed to meet the entered runway target. These are separate conditional alternatives, not recommendations; already-met, no-solution and out-of-horizon states are explicit.

All current models use user-entered assumptions or illustrative defaults, not industry benchmarks. Prefer official or institutional sources, then academic or textbook explanations; Variare’s implementation-specific formulas and conventions are disclosed separately. Sources support definitions or context without supplying live data or endorsing the model. No current model loads live market data or historical CPI data, and competitors are not mathematical authorities.

Uncertainty and omitted variables belong with the result. Constant-rate and simplified cost models cannot capture every change in prices, returns, demand, or operating conditions. Hiring’s cash projection omits working-capital timing and financing; negative cash is an unfunded gap, not proof that operations can continue. A precise number does not make an assumption certain or an output a recommendation.

Methodologies may evolve. Material changes should be documented with the affected model version and a real review date; each model retains its own review record. No AI model calculates user results and the application makes no AI API calls. We do not claim independent professional certification.

Platform methodology and editorial standards version 0.3 · Reviewed October 3, 2026. This review reconciles the guidance with three Explorers and the public Hiring R&D prototype; their engines and individual mathematical review dates are unchanged. For calculation, methodology or source corrections, contact [email protected].

Explore all ↗