Seasonality

Seasonality is the recurring, calendar-driven pattern in demand. Christmas trees in December. Sun cream in July. Heating pads in January. Seasonality is the part of the demand curve you could draw on a calendar, separate from the underlying trend (long-term growth or decline) and the random noise.

A clean planning process treats seasonality as a known — something to forecast against, not a surprise. The brands that get caught out aren't usually surprised by Christmas; they're surprised by when Christmas demand actually arrives, and by how much it's amplified by promotional pull-forward.

How seasonality is measured

The most common technique is seasonal indices — numbers that say "this month runs at 130% of average" or "this week typically sells at 60% of average."

Build them with three steps:

1. Take 2–3 years of weekly or monthly demand for the SKU.
2. Calculate the moving average to remove trend.
3. For each period, divide actual demand by the de-trended average. Average those ratios across years.

The result is a 12-number index (for monthly) or 52-number index (for weekly) that captures the seasonal shape. Multiply your underlying forecast by the index for each period to get the seasonal forecast.

Different timescales of seasonality

Seasonality isn't just annual:

• Annual — Christmas, Black Friday, summer holidays, school terms, back-to-school
• Monthly — payday spikes (mid- and end-of-month for D2C; some buyers cluster orders around month-end)
• Weekly — most consumer brands sell more Mon-Wed than Sat-Sun (or vice versa, depending on category)
• Daily — peak hours in the day

For most consumer brands, annual + weekly is enough. Monthly matters more for B2B. Daily matters mostly for staffing and ad spend, less for inventory.

Why seasonality is the easy part

Seasonality is predictable. That's the whole point. The hard parts of demand planning are the things that aren't seasonal — promotional spikes, viral moments, channel changes, competitor failures. A forecast that handles seasonality well still has to handle the rest.

But seasonality is the gateway: if your forecast doesn't handle the cycles you can literally circle on a calendar, it isn't going to handle anything harder.

Seasonality vs trend vs noise

Three components of any demand series:

• Trend — the underlying growth or decline rate
• Seasonality — the calendar-driven cyclical pattern
• Noise — everything else (the residual after removing trend and seasonality)

A SKU growing 20% YoY with a strong Christmas seasonal pattern will look chaotic month-to-month if you don't separate the components. Separate them and the picture becomes much clearer.

How seasonality interacts with planning

Three operational implications:

• Lead-time exposure. A SKU with a 14-week lead time and an annual seasonal peak means you have to commit to peak inventory in spring. That decision is locked in before the season starts. Get it wrong and there's no recovery.
• Safety stock during peaks. Demand variability usually rises during peak periods — peak weeks are noisier as well as bigger. Safety stock has to scale with that, not just the mean.
• Working capital cycle. Seasonal businesses have predictable cash-flow stress points. The buying happens months before the revenue. Treasury planning has to account for it.

Common seasonality mistakes

Three pitfalls:

• Treating last year as truth. If last year's peak was distorted by a promotion, supply issue, or one-off event, baking it into the seasonal index propagates the distortion forward.
• Conflating seasonality with promotion. A November spike in sales is mostly Black Friday, not "natural" November demand. Strip the promotional impact before computing the seasonal index.
• Same seasonality for all SKUs in a category. Some categories have huge cross-SKU variation in seasonality — best to compute per SKU where possible, with a category-level fallback for low-data SKUs.