The short answer
Seasonal ROAS compares revenue and advertising cost during a period when demand, prices, or competition may differ from normal. A holiday or promotion result can be useful, but it is not automatically a reliable forecast for an ordinary week. Keep the calendar context and the economics of the offer visible.
Choose a comparison that answers your question
A previous week comparison is quick, but may mix different weekdays or promotional conditions. A previous-year comparison may align the season while introducing changes in products, prices, tracking, and the business itself. Document those differences rather than treating either comparison as a perfect control.
Separate revenue return from contribution
Illustrative normal period: 100 orders at $100 generate $10,000 revenue with $2,000 advertising spend, or 5× ROAS. Assume $60 variable cost per order; contribution after ads is $10,000 − $6,000 − $2,000 = $2,000.
Illustrative promotion: 150 orders at a discounted $80 generate $12,000 revenue with the same $2,000 spend, or 6× ROAS. If per-order variable cost remains $60, contribution after ads is $12,000 − $9,000 − $2,000 = $1,000. More orders and a higher revenue ratio leave less contribution under these assumptions.
Let outcomes mature
Promotional orders may have different fulfillment, cancellation, or return timing. Compare equally mature cohorts where practical and mark provisional figures. Do not compare an early gross-sales number with an older report already reduced for refunds.
Plan with scenarios, then review
Write down assumed click cost, conversion rate, order value, and relevant costs before the season begins. Use more than one scenario if those inputs are uncertain. These are planning assumptions, not promised outcomes.
After the period, compare observed inputs with the assumptions and record what changed. If you need to know whether a promotion caused extra purchases rather than shifting their timing, a simple before-and-after ROAS report is insufficient; use a comparison design that addresses that question.
Separate calendar and offer effects
- Gather: Record promotion dates, discount economics, weekday mix and comparable mature outcomes.
- Produce: Compare contribution as well as ROAS, and retain differences from the comparison period.
Decision boundary: A holiday result is not a reliable ordinary-week forecast without supporting evidence.
Next reading
Examples are illustrative, not client results or benchmarks. The calculations use the inputs shown; real-world decisions depend on your records and assumptions. Report a correction.