Conversion Rate Traffic-Mix Bridge Calculator

What you can calculate

Enter counts for two mutually exclusive, exhaustive groups in two periods. A converting session has at least one qualifying conversion and is counted once, regardless of how many orders or events it produced. Both periods must use the same source definition and sufficient observation time. This local calculator does not access analytics accounts or estimate causal lift.

Every group must have at least one observed session in both periods. Zero converting sessions is valid; zero sessions leaves that group's rate unknown. The tool rejects empty, negative, non-integer, non-finite and oversized inputs and converting counts above their own denominator. Each input is capped at one billion to bound numeric operations. That technical cap is not a recommended sample size.

The exact descriptive identity

For group i and period t, let n be sessions, c converting sessions, r = c/n and w = n divided by all sessions in that period. The aggregate rate R = Σ(w × r). Rates and weights are fractions inside the calculation.

• Observed change ΔR = R1 − R0.
• Mix component M = Σ[(w1 − w0) × (r1 + r0)/2].
• Within-group component W = Σ[(r1 − r0) × (w1 + w0)/2].
• M + W = ΔR, up to floating-point rounding.
• Comparison at baseline mix S = Σ(w0 × r1).
• Fixed-baseline change = S − R0. This is not an additional component to add to M + W.

Multiply a rate difference by 100 to report percentage points. Relative change is (R1/R0 − 1) × 100%, undefined when R0 is zero. The symmetric identity splits the interaction evenly by averaging the two possible ordering paths; it is an arithmetic convention, not a causal allocation.

Check the synthetic default

Baseline A: 160/8,000 = 2%; B: 200/2,000 = 10%. Comparison A: 60/2,000 = 3%; B: 880/8,000 = 11%. Baseline weights are 80% A and 20% B; comparison weights reverse. Aggregate rates are 3.6% and 9.4%. Change is +5.8 pp, made up of +4.8 pp mix and +1.0 pp within-group movement. Relative change is approximately 161.1111%, not 5.8%.

Holding old weights fixed gives 80% × 3% + 20% × 11% = 4.6%, a +1.0 pp standardized change. It happens to equal the symmetric within-group component in this example because both group rates change equally. That equality does not hold generally.

Interpretation and boundaries

A negative mix component can offset improving group rates; conversely a positive headline can coexist with declining rates in both groups. Neither scenario reveals why the rates moved. Group selection, tracking changes, seasonality and unobserved intent remain possible explanations. This is not a significance test, confidence interval, revenue model or media optimizer.

The calculation preserves input counts and full-precision output in its export. Labels are local and not sent to analytics; do not paste personal data. Changing an input clears stale results. Use Reset example to recover the documented fixture. One group's zero numerator is supported; a zero overall baseline produces an undefined relative percentage but valid percentage-point arithmetic.

Do not bridge incompatible baselines

Shopify explicitly identifies September 21, 2026 as a new session-based baseline. If a comparison crosses a changed definition, this calculator can still compute numbers but cannot make them like-for-like. Read the measurement notice. Document continuity before using the result to guide a growth decision.