Discretionary Spending Ecommerce: How to Read Demand

Every example and amount on this page is illustrative. It shows the method and is not a benchmark, a quote or advice.

What Discretionary Spending Ecommerce Actually Measures

Discretionary spending ecommerce refers to online demand for purchases households can delay, reduce or skip when budgets tighten. Common examples include apparel, electronics, home goods, furniture, beauty products and jewelry. The boundary is not absolute: a coat may be essential for one household, while a premium accessory is discretionary for another. The useful question is whether the purchase is flexible, not whether the product has a fixed label.

Consumer spending is the broadest measure. It describes what households spend across the economy, including purchases from physical stores, marketplaces, direct-to-consumer sites and service providers. It can show whether consumers are allocating more or less money to goods and services overall, but it cannot tell you whether your store captured that demand or whether your margins were healthy. Reading US consumer spending for ecommerce provides the broader context.

A merchant’s sales data answers a narrower question: what happened inside the store. Sales should be based on completed, delivered orders, with cancellations and returns handled separately. Conversion rate measures the share of visitors who purchased; it does not measure demand among people who never reached the site. Product mix shows where revenue or units came from, so a stable total can hide customers shifting from high-priced electronics to lower-priced accessories.

Use each measure for a different decision. Consumer spending helps frame market conditions. Sales and conversion help diagnose store performance. Product mix, returns, and profitability help determine whether a category is worth supporting. ORVX connects Shopify orders, products and customers, then shows product- and collection-level profitability rather than treating revenue as demand quality. Its retail spending guide explains how to connect the market view to store-level evidence.

A Plain-English Formula for Demand Pressure

Two measures answer different questions. Discretionary sales share shows how much of your store’s sales comes from purchases customers can postpone: discretionary sales share = discretionary sales ÷ total sales. If discretionary products generated $40,000 of $100,000 in realized sales, the share would be 40%. Use this measure to understand your exposure: a store with a larger discretionary share may be more sensitive to changes in household budgets than a store whose sales are concentrated in essential products.

Demand change shows whether sales are rising or falling against a comparable baseline: demand change = (current-period sales − baseline-period sales) ÷ baseline-period sales. The baseline should be like for like: compare the same number of days, the same product or category scope, and the same sales channel. For seasonal ecommerce, that may mean comparing a holiday week with the equivalent week last year rather than with the previous week. Keep the currency consistent and decide whether the measure uses orders, units or realized revenue before comparing periods.

Use discretionary sales share when the decision concerns exposure or mix, for example, whether a product portfolio depends heavily on delayable purchases. Use demand change when the decision concerns momentum, for example, whether a category is weakening relative to a comparable period and deserves closer review. They can move in opposite directions: discretionary sales may account for a larger share because essential sales fell, even while discretionary demand also declined. Reading both measures together gives a clearer view of consumer pressure. For broader context on interpreting spending signals, see Reading US Consumer Spending for Ecommerce.

A Worked Example With Real Store Numbers

Consider a store selling discretionary home goods. In the prior year, it recorded 10,000 sessions, a 2.5% conversion rate and an $80 average order value. That produced 250 orders and $20,000 in delivered sales. The $80 average order value came from 300 units across those orders, or 1.2 units per order, at an average realized price of $66.67 per unit.

This year, delivered sales are down 12%, to $17,600. The store received 9,000 sessions, converted 2.4% of them, and produced 216 orders. Its average order value is now $81.48, which might initially suggest that demand is holding up. The unit detail tells a different story: customers bought 275 units, or 1.27 units per order, at an average realized price of $64. The higher units per order partly offset weaker traffic, conversion and realized price or product mix.

Read the movement as a chain rather than as an AOV result: sessions × conversion × units per order × realized price. Traffic fell 10%, and conversion fell from 2.5% to 2.4%, reducing orders from 250 to 216. Units per order rose by about 5.8%, while realized price fell by about 4%. Multiplying the year-over-year factors gives approximately 88% of the prior sales level, consistent with $17,600 versus $20,000 after rounding.

That decomposition points to different next checks. A traffic problem calls for a channel or campaign breakdown; a conversion problem calls for product, device, landing-page or checkout cuts; and a realized-price decline calls for product-mix, discount and return analysis. The higher basket quantity may reflect bundling or a shift toward lower-priced items, not stronger willingness to spend. AOV alone cannot distinguish those explanations, and it does not show whether the extra units were profitable. The next decision therefore depends on margin and contribution by product or collection, not merely on recovering the headline sales figure.

• Prior year: 10,000 sessions × 2.5% conversion = 250

The formula, stated plainly

The calculation is only as useful as the inputs behind it. Set a fixed period, such as a calendar month, and use the same period for the comparison. Do not compare a 31-day month with a 20-day promotional period without labeling the difference. For a seasonal store, compare with the same month last year or with a defined trailing-period baseline rather than an arbitrary prior month.

Define the discretionary product set before calculating anything. Include products customers can reasonably postpone, and keep the classification stable across both periods. If a product moves between collections or categories, preserve the rule used to classify it; otherwise, a change in the product mix can look like a change in consumer demand.

Use realized revenue for the sales calculation: delivered orders that remain completed, excluding cancelled and returned orders. Gross sales, captured payment totals or shipped-order revenue can produce different answers. The denominator must use the same revenue definition as the discretionary numerator and cover the full store, not just the products included in a campaign.

Treat the two outputs as different decision signals. The discretionary share tells you exposure: how much of the store depends on purchases customers may defer. The period-over-period change tells you direction: whether that discretionary demand is expanding or contracting against the selected baseline. A high share with stable demand calls for monitoring concentration and margin. A falling change in demand calls for examining traffic, conversion, average order value, product mix and contribution after advertising before cutting spend.

Revenue alone can hide what changed. A decline may come from fewer orders, lower prices or a smaller basket rather than fewer shoppers. Read the demand calculation alongside sessions, orders, conversion rate and average order value, while keeping the core formula unchanged. That separates a customer-demand signal from a merchandising or acquisition effect.

• Use one currency, one time basis and one revenue definition throughout.
• Freeze the discretionary product classification before comparing periods.
• Act on exposure when the share is the concern; investigate movement when the change rate is the concern.
• Use realized revenue, not booked revenue that still includes cancellations or returns.

A worked example with real numbers

Take an illustrative Shopify home-goods store comparing September with the same month last year. In the prior September, it had 40,000 sessions, a 2.5% conversion rate and a $72 average order value. That produced 1,000 orders and $72,000 in realized discretionary revenue. In the current September, traffic rose to 42,000 sessions, but conversion fell to 2.0% and average order value fell to $68.57. The store therefore had 840 orders and $57,600 in realized discretionary revenue.

The arithmetic shows two different signals. Discretionary demand fell by $14,400, or 20%: ($57,600 − $72,000) ÷ $72,000. If total realized store revenue fell from $100,000 to $96,000, discretionary sales share also fell, from 72% to 60%. That share says the store is now less exposed to postponable purchases. The 20% change says customers are buying materially less from that category. The first measure is useful for assessing revenue mix; the second is the one to act on when deciding whether demand itself has weakened.

Profitability changes the decision again. Suppose current discretionary cost of goods was $23,000 and advertising was $15,000. Contribution after advertising was $19,600 ($57,600 − $23,000 − $15,000), or 34.0% of revenue. In the prior year, $25,200 of product cost and $12,000 of advertising produced $34,800, or 48.3%. Revenue declined 20%, but contribution declined 43.7%. Cutting spend solely because sales are down could miss a collection with worsening economics; scaling an apparently stable category could compound the problem.

The mistakes that make the number wrong

The first error is using a sales number that is not comparable. Gross order value can include cancelled orders, returned items, taxes, shipping or discounts depending on the report. A store may appear to have stable discretionary demand while refunds are rising underneath it. Use one revenue definition for both periods, and document whether the denominator is orders, units, sessions or realized revenue. ORVX calculates realized revenue from delivered orders only, excluding returned and cancelled orders.

Timing creates a second false signal. Comparing a promotional week with an ordinary week, or a partial current month with a complete prior month, makes demand pressure look larger than it is. Align calendar periods and account for shipping cutoffs, major launches, paydays and promotional events. Year-over-year comparisons reduce some seasonality, but they do not remove the effect of a different discount schedule or product availability.

Category boundaries can also corrupt the result. A merchant may classify a replacement filter as discretionary because it sits in a home-goods collection, while customers treat it as a necessary replenishment. Conversely, an expensive upgrade may be grouped with essential products and hide a pullback in postponable demand. Set the classification at the product or use-case level, then keep it unchanged across the comparison. Do not combine revenue share with unit demand or conversion rate and describe all three as the same measure.

Returns, discounts and marketing mix need separate checks. A higher order count can coexist with lower realized revenue if average discounts and refund value increased. A lower attributed ROAS does not automatically mean consumers stopped buying; it may reflect channel allocation, attribution rules or rising acquisition costs. MER and attributed ROAS answer different marketing questions and should not be substituted for a demand measure. For a broader guide to interpreting the signals, see Reading US Consumer Spending for Ecommerce.

Before acting, reconcile the store report to the inputs behind the comparison. Confirm the same trading currency, product set, traffic scope and order status in both periods. Check inventory and dead stock as well: a sales decline caused by an out-of-stock product is a supply constraint, not proof of weaker discretionary spending. ORVX connects to Shopify and advertising accounts, reports return rate by units and refund value against revenue, and keeps each merchant's trading currency separate. Those checks make the number more defensible; they do not turn a noisy metric into a forecast of what customers will do next.

• Use realized revenue, not gross checkout value, when returns or cancellations are material.
• Match periods, promotions, product availability and currency before comparing demand.
• Separate demand measures from profitability and marketing measures: revenue, gross profit, contribution after advertising, MER and attributed ROAS answer different questions.
• Investigate stockouts, assortment changes and classification errors before interpreting a decline as weaker consumer demand.