Definition
Seasonality is the predictable, recurring variation in demand linked to the calendar. It operates at several scales at once: across the year (months and seasons), around fixed events (holidays, festivals, school terms, financial year ends), across the week (weekday against weekend) and across the day (morning against evening). What makes it seasonality rather than a trend is repetition: the same shape returns each cycle, even if the overall level rises or falls.
Two kinds matter in marketing. Demand seasonality describes when people are in the market. Media seasonality describes when attention is expensive, because competitors share the same calendar. The media planning framework points out that the two do not always coincide, and that buying reach in the cheaper weeks before a peak can beat spending everything at the peak price.
Reading it from your own data
The most reliable picture comes from the business's own records over at least a full year, and ideally two or more: sales, enquiries or sign-ups by week or month, site traffic by source, and Search VolumeThe estimated number of times a query is searched in a period, usually per month. for the category's main terms from keyword tools and Search Console. Plotting them together shows the shape of the year, the troughs, and the lead time between interest rising and orders arriving, which sets when awareness work has to start. The annual planning guide treats this as part of the yearly situation review, and suggests that a new business borrows the category's search seasonality until it has a year of its own data.
Two cautions apply. Events whose dates move need aligning by event rather than by calendar month, and a single year can contain one-off shocks, such as a heatwave, that look seasonal but will not repeat.
In practice: phasing budgets, bids and calendars
Once the shape is known, it enters plans in three places.
Budgets. An annual budget split evenly by twelve starves the peaks and wastes money in the troughs. Phasing the budget by month in proportion to expected demand, and planning peak spend on the cost seen in the same weeks last year rather than on the yearly average, keeps money available when it is needed. The seasonal peak walk-through for the marketing budget calculator shows how to plan a peak's total that way.
Bids and pacing. Within a platform, a budget raised before a known peak is worth more than one raised after the campaign has already been limited. Smart BiddingGoogle Ads' automated bid strategies that set each auction's bid from conversion signals. learns from recent conversion rates, so a short, sharp event can catch it out; Google documents seasonality adjustments for such events, and the budgets and pacing guide explains why they change bids but not budget, so both need setting. Day-to-day Budget PacingHow fast a campaign spends its budget across the flight compared with the plan. then checks whether spend is tracking the seasonal curve rather than a flat line.
Calendars. Peaks are fixed dates that the rest of the plan fits around. The campaign calendar guide places them first and plans backwards from each one, so that media, creative and content are ready before demand rises. Where the category has sharp seasons, FlightingScheduling advertising in bursts separated by periods of little or no activity. concentrates spend into those windows and accepts silence between them.
Seasonality in reporting and tests
The most common reporting error is comparing a month with the one before it. In a seasonal business that comparison mostly measures the calendar. The fairer baseline is the same period a year earlier, adjusted for known changes in prices, range or spend, or a trailing period with the seasonal pattern stated. The SEO KPIs and reporting guide sets organic baselines that way, because traffic totals rise and fall with the season whatever the SEO work does.
The same applies to tests and budget experiments. A change launched into a rising month looks better than it is; one launched into a falling month looks worse. Randomised tests that run both arms at once avoid the problem, while before-and-after comparisons need the seasonal shape taken out first.
A ski hire shop changes its comparison
Snowcrag Ski Hire ran search and social campaigns through October and November and reported a success: online bookings were 340, against 95 in August and September. Compared with the same 8 weeks a year earlier, though, bookings were 352, so the campaigns had added nothing measurable. Two years of data showed that searches for ski hire in the area barely moved until the first forecast of snow, then rose sharply over about 3 weeks. The next winter the shop cut its October budget from £3,200 to £1,100, held the rest in reserve, and released it when search volume began to climb. Bookings for the season rose from 1,180 to 1,310 on the same total spend of £9,000.
Related terms
Budget PacingHow fast a campaign spends its budget across the flight compared with the plan. keeps daily spend in line with a seasonal plan. FlightingScheduling advertising in bursts separated by periods of little or no activity. is the media pattern that concentrates spend into seasonal windows. Smart BiddingGoogle Ads' automated bid strategies that set each auction's bid from conversion signals. needs warning of short, predictable events because it learns from recent data, and Search VolumeThe estimated number of times a query is searched in a period, usually per month. is usually the earliest, most public signal of a season starting.
Common confusions
- Seasonality is not performance. A rise that happens every year at this time is the calendar, not the campaign.
- Month-on-month is not a fair comparison in a seasonal business; the same period last year is.
- Demand peaks and cost peaks are not the same thing. Auction prices rise with competition, which may start earlier or later than buyer demand.
- Seasonality is not only annual. Weekday and hour-of-day patterns are seasonal too, and matter for scheduling and pacing.
- Industry calendars are not your calendar. The business's own data, read over more than one year, beats any general assumption about when people buy.
