Blog · Territory, capacity and quota planning
Odd even pricing compares price endings such as $39.99 and $40.00. Work through a controlled sales example, reconcile contribution, and see how to test the result on your own transaction export.
Odd even pricing uses price endings, such as $39.99 and $40.00, to change how customers perceive an offer. Odd pricing commonly means just-below prices, while even pricing commonly means rounded amounts. Test the effect on contribution as well as units: an ending that sells more does not automatically earn more.
A just-below price may affect the apparent price level or signal a bargain. A rounded amount may fit a different presentation of value. Those are hypotheses about your customers, rather than dependable outcomes for every product.
In their field experiments on price endings, Eric Anderson and Duncan Simester found higher demand for nine-ending prices in the catalog settings they studied. The research also found differences associated with item history and sale cues. That supports testing the actual offer in its selling context; it does not supply a universal uplift to put into a budget.
Changing $40.00 to $39.99 changes the amount and its presentation together. A comparison of those offers measures their combined commercial effect. It cannot, by itself, prove how much came from the ending rather than the lower price.
Decide the product, eligible customers, test period, and primary measure before the offer starts. Assign comparable customers to the offers concurrently, and keep a customer's assignment consistent. For an account-based seller, use the account as the assignment unit so different buyers at one customer do not receive conflicting offers.
Keep product availability, delivery promise, placement, and promotion consistent. Where stores or channels differ materially, randomize within comparable groups and examine those groups separately. NIST's guidance on randomized block designs explains how grouping comparable experimental units helps control known sources of variation. Applying that principle here, compare endings within channels before treating a channel mix change as a pricing result.
Keep two exports. The assignment file identifies each eligible customer, their group, the offer shown, and the date. The transaction file identifies order lines, group, product, units, realized price, discounts, refunds, and variable costs. Customers who buy nothing must remain in the exposure denominator.
For a transaction-level check of prices actually collected, use price realization by customer. List prices alone do not establish what the business earned.
Consider an illustrative US office-supply seller testing one accessory. Each assigned shopper can buy one unit during the test. Both groups have the same eligible exposure, no discounts or returns, and the same variable cost per unit, including product and variable fulfillment costs. Sales tax collected for authorities is excluded from revenue.
| Test input | Even ending | Odd ending |
|---|---|---|
| Eligible shoppers | 10,000 | 10,000 |
| Units sold | 1,000 | 1,060 |
| Realized unit price (USD) | 40.00 | 39.99 |
| Variable cost per unit (USD) | 25.00 | 25.00 |
Use the same definitions for both groups:
Net revenue = Units sold × Realized unit price
Contribution = Net revenue − Variable costs
Contribution per eligible shopper = Contribution / Eligible shoppers
The calculations below retain cents for money. Rates are displayed to two decimal places where needed, and contribution per shopper to five decimal places. The underlying calculations use unrounded values.
| Calculation | Even ending | Odd ending |
|---|---|---|
| Buying rate: units / eligible shoppers | 10.00% | 10.60% |
| Net revenue (USD) | 1,000 × 40.00 = 40,000.00 | 1,060 × 39.99 = 42,389.40 |
| Variable costs (USD) | 1,000 × 25.00 = 25,000.00 | 1,060 × 25.00 = 26,500.00 |
| Contribution (USD) | 40,000.00 − 25,000.00 = 15,000.00 | 42,389.40 − 26,500.00 = 15,889.40 |
| Contribution per eligible shopper (USD) | 15,000.00 / 10,000 = 1.50000 | 15,889.40 / 10,000 = 1.58894 |
The odd offer sells 60 more units: 1,060 − 1,000. Its relative unit increase is 1,060 / 1,000 − 1 = 6.00%. The buying-rate change is 10.60% − 10.00% = 0.60 percentage points; those descriptions are not interchangeable.
Revenue increases by $2,389.40, calculated as $42,389.40 − $40,000.00. Additional variable costs are 60 × $25.00 = $1,500.00. The contribution gain is therefore $889.40, or $2,389.40 − $1,500.00. Relative contribution increases by $889.40 / $15,000.00 = 5.93%, rounded.
For the distinction between variable costs and the costs included in gross profit, see gross margin vs contribution margin vs net margin.
Reconcile the groups to the filtered transaction export. Combined revenue must be $40,000.00 + $42,389.40 = $82,389.40. Combined variable costs must be $25,000.00 + $26,500.00 = $51,500.00. The difference, $30,889.40, must equal $15,000.00 + $15,889.40 of group contribution.
Now change only the odd group's units to 1,000. Keep its price, unit cost, and eligible exposure unchanged.
| Variation calculation | Odd ending, unchanged demand |
|---|---|
| Units sold | 1,000 |
| Net revenue (USD) | 1,000 × 39.99 = 39,990.00 |
| Variable costs (USD) | 1,000 × 25.00 = 25,000.00 |
| Contribution (USD) | 39,990.00 − 25,000.00 = 14,990.00 |
| Contribution change against even group (USD) | 14,990.00 − 15,000.00 = −10.00 |
Without a demand change, the lower price loses $10.00. To match the even group's contribution, the odd group needs 15,000 / (39.99 − 25.00) = 1,000.66711 units, rounded to five decimals. Because units are whole, it needs at least 1,001 sales. This is the commercial break-even point under the example's fixed unit-cost assumption; it does not establish statistical confidence in an observed difference.
The broader method for costs and required volume belongs in break-even analysis.
The pricing owner decides whether to extend the offer. Finance checks the contribution definition and export reconciliation. Operations checks stock availability and whether the groups actually received the intended prices.
Before adopting the observed winner, review uncertainty using the original assignment unit. Repeated purchases by one account are not independent new customers. Check whether the result persists across the groups specified before the test, and allow the agreed return window to close. The worked totals establish arithmetic, not evidence that the effect will repeat.
Record the offer, cost policy, observed result, and decision together. That gives the next review a comparable starting point instead of an unexplained new price list.
Bring your assignment and sales exports with group, product, units, prices, returns, and variable costs. Covirage maps the columns, and its tools compute and check the supplied revenue and contribution figures; the external AI model explains the result with every figure traced to its rows. Create a free Covirage account and bring a sample of your own data.
See revenue driver analysis for the supporting business context.
No. The effect depends on the product, customer, channel, and surrounding offer. A result from another retailer is a reason to test an ending, not a forecast for your business. Keep availability, promotion, and measurement consistent across the comparison.
A rounded price of $40.00 is an example of even pricing. The relevant distinction is the presentation of a round amount, rather than whether the last digit is mathematically even. Compare it with the specific alternative your team is considering.
Measure both, but choose the decision measure before starting. Contribution after variable costs shows whether added sales pay for the price change. Revenue can increase while contribution declines if the additional units carry substantial fulfillment, return, or acquisition costs.
The pricing or merchandising owner should define the offer and decision. Finance should reconcile realized revenue and variable costs. Sales operations or ecommerce operations should preserve group assignment, exposure, availability, and promotion records so the comparison can be explained.