Methodology

How the calculations work

Calculation engine v0.1.0 · last updated 2026-10-16

Principles

• Deterministic: the same inputs always give the same result. No AI model computes any number.
• Exact decimal arithmetic for money, never binary floating point. Rounding happens once, on the payout, half-up to cents.
• Nothing is inferred: quota, currency conversion, attainment, clawback eligibility and dates are never guessed. Missing or contradictory inputs produce an explicit error or warning.
• Every number is explained: each step appears as a line item, and line items always add up to the payout.
• Local processing: calculations run in your browser; share links contain plan settings only.

Calculation rules

Tier boundaries

Each tier is the half-open interval [from, to): the lower bound is included, the upper bound belongs to the next tier. Tiers must start at 0, be contiguous, and only the last tier may be open-ended.

Marginal vs retroactive

Marginal (graduated): Σ (revenue within each tier × that tier's rate). Retroactive: all revenue × the rate of the tier the total falls into. Retroactive ladders create cliffs at each threshold; the calculator always shows both results.

Tiers in % of quota

Thresholds expressed in % of quota are converted with the quota you enter. Without a quota, the calculation stops with an error.

Accelerators

Incremental (default): revenue above the threshold × tier rate × (multiplier − 1) is added to the base commission. Retroactive: once the threshold is reached, all commission is multiplied.

Order of operations

Credit (splits) → tiers → accelerator → cap → clawbacks → guaranteed minimum → rounding. Clawbacks can make a payout negative when no minimum applies; the result says so rather than hiding it.

Splits

Commission is computed on the credited share. Shares that do not add up to 100% trigger a warning and are not silently rescaled.

Team simulation

Synthetic attainment is drawn from a normal or uniform distribution with a fixed seed (or taken from your list), rounded to 0.01% of quota, floored at 0. Both plans are applied to the same draw. These are modelling assumptions, not market benchmarks.

Testing

The engine is covered by automated tests: canonical cases (zero revenue, exact thresholds, one cent above and below, 100% and above quota, caps, splits, clawbacks, rounding, duplicate IDs, malformed numbers, currency mismatches) and property tests over thousands of random plans (line items reconcile to the payout, caps are never exceeded, no NaN or infinity, determinism, marginal ≤ retroactive for increasing rates).

Hand-calculated worked examples are asserted against the engine on every build.

Review status

Not yet independently reviewed. The rules and worked examples were written and tested by the developer. An independent review by a sales compensation practitioner is planned before launch; until then, treat every result as an estimate and check it against your plan document.

Scope and limits

• Not modelled yet: draws, ramps, territory and team overrides, multi-period true-ups, bookings vs collections timing, currency conversion.
• US-centric explanations. Compensation rules and labour law differ by country (France, Spain and Brazil included); nothing here is legal, payroll or tax advice.
• Software suggestions use fixed thresholds described on the software page; they do not name vendors.