Measurement / calculator / Free to use

Free Shipping Threshold Replay and Recovery Calculator

Replay two shipping thresholds on the same order bands, quantify existing-order subsidy and calculate the additional-order hurdle without inventing conversion lift.

Growthcraft Editorial · 2026-10-03. AI-assisted research and implementation. Examples are synthetic; Akshay's personal review is not claimed.

Enable JavaScript to change inputs in the interactive calculator. The complete formulas, default example and limitations are available below.

A fixed population rather than an uplift forecast

Enter one period of eligible orders in a single currency. Each line contains merchandise basket value, variable cost excluding carrier expense, carrier expense and number of identical orders. Basket value is after discounts, before tax and shipping. Variable cost includes the product and other applicable marginal costs, but not the carrier expense entered separately. All amounts exclude tax.

The model charges one flat shipping fee below each threshold and zero at or above it. Every order incurs the entered carrier expense in both policies. Exclude or separately model zones and goods with different rules. A band is exact only when its orders share the entered amounts; do not average together baskets that fall on different sides of either threshold.

Equations in the same currency

For order band i, contribution under threshold T is count × [basket − variable cost − carrier cost + fee × indicator(basket < T)]. Sum across bands for each policy. Contribution change is proposed minus current. A positive number is a static gain, not proof that fewer customers will abandon.

  • Newly free-shipping orders previously paid the fee but now qualify.
  • Orders losing eligibility previously qualified but now pay the fee.
  • With unchanged baskets and costs, contribution change = fee × (losing eligibility − newly qualifying).
  • Loss to recover = max(0, −contribution change).
  • Additional orders needed = ceiling(loss / net contribution per genuinely additional order).

Additional-order contribution must reflect the proposed policy and all marginal costs, including acquisition cost where relevant. The tool accepts a nonnegative value: zero with a positive loss yields an unattainable hurdle, not division by zero. If new orders are loss-making, this recovery method is unsuitable; adding them cannot offset the loss.

Check the default

Synthetic, currency units. Current threshold 75, proposed 50, fee 5. The rows are 40,24,6,100; 60,36,6,80; and 90,54,7,40. Current contribution is 4,500; proposed contribution 4,100. Free-shipping orders rise from 40 to 120, giving 80 newly subsidized orders and a 400 static loss.

At 12 contribution per genuinely additional order, 33 orders contribute 396 and do not quite recover the loss; 34 contribute 408 and do. The required 34 equals 15.45% of 220 historical orders. This percentage is a count comparison, not an estimate of conversion-rate lift, statistical significance or causality.

Validation and precision

Amounts accept nonnegative decimal strings with at most two decimal places and a maximum of 1,000,000 per field. Arithmetic uses integer hundredths. This implementation is intended for currencies modeled with two decimal places; adapt before using currencies with another minor-unit convention. Counts are positive whole numbers, with at most 200 bands and one million total orders. These are technical safety limits, not recommended business volumes.

Blank rows, headers, symbols, negative values, non-finite amounts, fractional counts and extra columns are rejected. Negative order contribution is permitted when valid nonnegative costs exceed revenue. A blank export is missing data, not a zero-order scenario. A threshold of zero makes all nonnegative baskets eligible. Equal thresholds or a zero fee produce zero replay change.

Use the result carefully

The replay omits behavior, changes in basket composition, shipping cost jumps, split orders, cannibalized future orders and returns beyond your supplied cost convention. It does not search for an optimum threshold. If you lower the threshold and see no static loss, inspect whether any observed orders actually fall in the affected interval before concluding the change is free.

Copy/download preserves inputs, assumptions and exact integer-cent results. The scope label stays local and is not sent to analytics. Editing clears stale results. Use the framework to document evidence and the article for an independent executable replay. No login, paid API or audit allowance is needed.

Companion resources

Back to strategic resources