Should a seasonal revenue peak grow as the business grows?

Knowing that revenue peaks every December does not establish how large next December's uplift should be. A business that doubles its normal sales level might keep the same seasonal dollar increase, or it might experience a seasonal increase that doubles with the base. Those assumptions can produce different staffing, inventory and cash requirements.

Hyndman and Athanasopoulos distinguish additive patterns, where components are added, from multiplicative patterns, where seasonal variation scales with the level. The relevant question for finance is whether the observed seasonal effect is approximately stable in dollars or in proportion to the business's underlying size.

Start with a consistent revenue measure and several comparable cycles where available. A changed billing policy or new business unit can alter the apparent pattern. Do not treat every difference between corresponding months as seasonality before checking what the numbers represent.

Suppose an invented business has an underlying monthly revenue level of $100,000. A peak-season month is $140,000 and a quiet-season month is $60,000. At that level, the observations can be described either as seasonal effects of plus and minus $40,000 or as factors of 1.4 and 0.6.

Now suppose the underlying level becomes $200,000. Keeping the dollar effects unchanged gives a $240,000 peak and a $160,000 quiet month. Keeping the proportional factors unchanged gives a $280,000 peak and a $120,000 quiet month.

Both approaches match the original illustrative pair. They diverge by $40,000 in each new seasonal extreme. Evidence about how the pattern behaves at different business levels is needed to choose between them; the original two numbers alone cannot identify the scaling relationship.

Plot revenue over time and compare seasonal amplitudes at different underlying levels. A roughly constant dollar gap suggests an additive representation may be useful. A gap that grows in proportion to the level suggests a multiplicative representation may deserve testing.

Avoid inferring the relationship from one unusual peak. A promotion, supply shortage or unusually large customer order can distort a particular year. Review the operating evidence and assess whether the event is part of a recurring seasonal mechanism or a separate occurrence.

Where the company changed product or customer mix, examine whether the aggregate pattern combines different seasonal businesses. A growing subscription component may dilute a holiday-driven retail peak. The company-wide seasonal factor can change even when each component's underlying behavior is stable.

Check whether delivery capacity capped the recorded peak. A flat seasonal dollar uplift can reflect a fully booked team rather than a stable ceiling on demand. If the business plans to remove that constraint, its old fulfilled-revenue pattern may not describe the additional opportunity. Keep supported demand evidence separate from the revenue the constrained operation could previously deliver.

Seasonal effects can also evolve gradually. Test whether a fixed historical pattern remains suitable after material channel or customer changes, while avoiding frequent revisions driven by a single noisy month.

A full additive seasonal pattern is ordinarily centered so its effects sum to zero over the cycle. A corresponding multiplicative pattern is normalized around an average factor of one. This keeps the seasonal component from quietly adding another growth assumption to the overall level.

The two-month example illustrates extremes only; it is not a complete annual seasonal profile. A real monthly model needs all twelve months and the relevant trend treatment. Reusing only peak-month adjustments while leaving every other month untouched can overstate annual revenue.

Check the reconciliation after applying seasonality. The annual total should reflect the model's level and trend assumptions consistently, rather than an unexplained uplift introduced by copied percentages. Preserve the seasonal factors and their normalization with the forecast version.

Monthly totals can vary because months contain different numbers of trading days, holidays or billing runs. A December effect estimated from raw totals may include those calendar differences. Decide whether to model them separately or retain them within a clearly defined seasonal treatment.

Moving events require particular care. A holiday or customer renewal cycle that shifts between reporting periods can make fixed month labels misleading. Use evidence about actual dates when the timing affects staffing or the cash plan.

Do not double count an adjustment. If the base revenue forecast already uses operating days and known promotional dates, an additional seasonal factor estimated from unadjusted historical totals may repeat part of the same effect. Explain which variation each component is intended to capture.

Multiplicative relationships are easier to interpret for positive series. Zero or negative revenue observations, including large net credit months, need careful treatment. A model that divides by a zero level or takes a logarithm of a nonpositive amount is not repaired by silently dropping those periods.

A log transformation can express multiplicative components additively on the transformed scale, but the return to dollar forecasts needs the appropriate treatment. Preserve the distinction between a transformed central estimate and the mean dollar outcome used in a financial plan.

Evaluate candidate models on subsequent or appropriately withheld periods. Compare peak and quiet months as well as the overall error. An apparently good average fit can conceal a recurring underforecast of the peak that determines the business's largest commitment.

Connect the forecast's seasonal size to the relevant cost and cash drivers. If the larger proportional peak requires extra capacity, include the timing and cost of obtaining it. If the quiet month becomes lower, examine the fixed commitments that remain during that period.

Keep the alternative seasonal representation as a targeted sensitivity when evidence remains limited. State what new observation would make the team reconsider the chosen form. The decision should show how much depends on peak growth, rather than burying that assumption in a generic annual growth percentage.