Ecommerce Plan: Do Retained Orders Fund the Next Stock and Ad Budget?

Increase an ecommerce advertising budget only after the original order cohort has absorbed its refunds, fulfillment costs and acquisition spending, and the next purchase fits the dated bank forecast. A high sales-to-ad-spend ratio does not establish either retained contribution or money available to buy stock.

Start with the customers and goods you actually retain

Choose a purchase cohort, such as orders first placed in one week, and follow it through delivery, refund and return inspection. Reconcile placed orders, fulfilled orders, kept purchases and returned units separately. An exchange, a partial refund and a fully refunded single-item order need different rows. Do not compare a mature older campaign with a new campaign whose return window is still open.

The NRF and Happy Returns research released in October 2025 estimated that 19.3% of online sales would be returned in 2025. Its merchant survey covered 358 ecommerce professionals at US merchants above $500 million in revenue. This is a dated sales-value estimate for a different business population, not a measured 2026 return rate, a share of orders or the assumption to put into every small store. Use your own category and customer cohorts to cost returns.

Reconcile one complete 100-order cohort

Original assumptions, in dollars: fulfill 100 single-item orders at $60 each after buying 100 identical items for $20 each. Twenty orders receive full refunds. Fifteen returned items are inspected as saleable; five are lost. Outbound shipping costs $5 for every fulfilled order, return handling $3 for every return, processor fees 3% of original sales and acquisition $12 per placed order. No fee refund, tax, additional overhead or other stock is assumed.

The $300 inventory is neither another loss nor cash already collected. It is usable stock in this specific case. If inspection instead finds damage or a different SKU cannot sell in the next campaign, revise the stock and buying plan rather than assuming all returns replace new purchases.

Set the acquisition limit on one consistent denominator

Original-sales ROAS is $6,000 divided by $1,200, or five times. Using retained sales gives four times. Both omit product and delivery costs. Before acquisition, selected contribution is $2,360, or $23.60 per placed order. Spending $24 on each of the same 100 orders would produce a $40 loss before overhead. This boundary holds customer quality and all other costs constant; it does not predict what a larger ad budget will acquire.

Separate acquisition per new customer from acquisition per order. A repeat order can add contribution without a new customer acquisition, but it still needs merchandise, delivery and service. Do not count repeat orders twice or remove refunded customers from the original spending denominator. Add fixed platform, customer-service and brand commitments before estimating whole-store profit.

Use opening bank cash and the settlement calendar

Add an explicitly assumed $4,000 unrestricted opening bank balance before the purchases. Pay inventory, advertising and outbound shipping on day 0. Receive the original sales net of processor fees on day 2, then pay all selected refunds and return handling on day 14. Assume the cohort is closed at that review date.

The $860 is the increase in bank cash, not the closing balance. The minimum displayed bank balance is $300. If settlement instead arrives on day 15 while refunds remain due on day 14, cash temporarily falls to negative $960. Final contribution and final bank cash are unchanged, but another $960 is needed to avoid a negative balance during that delay.

Price the next campaign before spending the apparent surplus

For a separate next-batch decision, assume the 15 saleable returns can supply identical items in the next campaign and all selected upfront payments fall before its receipts. Choose a $1,000 bank floor for this decision. It is an assumption, not proof that rent, tax or other business obligations are funded.

Neither proposed order count is a demand forecast or verified warehouse capacity. Obtain the actual supplier minimum, due date, pick/pack throughput and processor hold terms. A phased purchase or smaller campaign needs its own quantity and price; credit is not available merely because it would make the table work. Avoid reserving the same expected refund twice inside both an explicit refund payment and a buffer.

Review cohort contribution, inspected stock, return timing and the lowest bank day together each week. The separate Ecommerce Financial Model page describes the selected product. Its retained-order planning unit needs an explicit bridge to the placed-order cohort here; none of these figures is a delivered-workbook result.