Seasonality in advertising: how to plan your budget on a monthly basis

The monthly budget is planned ahead of time: preparation starts two to three weeks before the peak, in steps, not with a single raise on the first day of the season. Seasonality is the recurring pattern in demand and in the cost of impressions, read from your own data for at least 12 months back, not from general industry expectations. The reason for getting ahead is technical. Algorithms learn slowly, and any sharp change in budget puts the campaign back into the learning phase exactly when weak results cost the most. In the peak period competition pushes up the cost of impressions, so the structure, creatives and tracking must be ready before that, not during the peak.

My name is Hristiyan Panov — a performance marketing consultant with over 5 years’ experience and over €2 000 000 managed ad budget for more than 90 companies on Google and Meta. Below is how to read your own seasonality and what not to touch at the peak.

How do you read your own seasonality?

Not from general industry articles, but from your own data: revenue by month for the last two years, split into number of orders and average value. The two move differently — sometimes orders grow while the average value falls, and total revenue misleads.

The procedure is mechanical. Pull the monthly values into a table with columns for revenue, number of orders and average order, and next to each month write its ratio to the yearly average. That shows which months rise above the line and by how much, instead of relying on a feeling. Do it for the two years separately — a month that rises in only one of them was a one-off event, not a season.

Overlay the spend as well. If last year’s peak came with a doubled budget, that was not a season but an investment. Next to revenue, put ad spend by month and ROAS: a month is seasonal if ROAS holds or rises at an unchanged spend.

Split the data by channel and by new versus returning customers. The season for new customers often runs ahead of the revenue season.

Why does the cost of impressions rise in season?

Because everyone is advertising at the same time. The auction is the same, there are more participants and with bigger budgets. The cost per click and per thousand impressions rises without anything in your account having changed.

In the report this looks recognisable. The cost per thousand impressions rises, CTR and the order rate stay at their earlier levels, and CPA climbs proportionally. That is the auction. If, however, CTR falls while the cost per thousand impressions stays put, the cause is the creative, not the season. The two states differ in which metric moves first.

The check takes minutes: compare the cost per thousand impressions and CTR in the peak period with those before it, in the same ad set, with the audience and creative unchanged. Isolate one factor before you react.

This means your break-even point does not change, but it is harder to reach — not that the campaign has suddenly become bad. The answer lies in the margin and in the offer, not in switching off ad sets that work.

When is the budget raised?

In steps, starting two to three weeks before the start of the period. The reason is the learning phase — you want the campaign to be stable at the new level before the traffic arrives.

A step is about 20% of the daily ad budget, after which you wait for the result to settle and only then comes the next one. Make the change where the budget sits: with CBO the raise affects the whole campaign and the redistribution between ad sets, with ABO only the specific ad set. Do not touch the budget and the bidding strategy on the same day, because afterwards you cannot tell what caused what.

The same applies to the exit: bringing the budget down after the season is also done in steps, otherwise you restart the learning exactly when you should be collecting data for the next period.

Record the date and size of every step on a separate row. Next year that is a ready schedule, not a memory.

What should you prepare before the peak?

The list is closed before launch. At the peak, each of these tasks eats up more time than you have, and gets solved on the most expensive traffic of the year.

• Creatives — ready and approved in advance; approval drags exactly when you are in a hurry. Keep spare variants in the same format too, so you can swap a creative without touching the structure.
• Stock and feed — rejected items are fixed beforehand, not during. Reconcile prices, availability and titles, because they diverge most often exactly during a promotion.
• Tracking — reconciled with the real orders, so you can trust the numbers when reacting quickly. Compare the orders in the platform, in GA4 and in the accounts for the same period.
• Website and checkout — this is the last moment to touch them. After launch, every change is tested on the traffic that costs you the most.
• Audiences — updated lists, recent buyers excluded, lookalike audiences refreshed.

Close the preparation with a real test order and check that the event reaches the platform with the right value.

What should you NOT do in the peak period?

Do not change the bidding strategy, do not restructure campaigns, do not launch a new checkout and do not change contractor. Each of these puts the system back into learning or carries a risk whose cost at the peak is many times higher.

In season, one thing is done: monitoring and budget adjustments within the permitted steps. The monitoring has substance — pace of spend against the plan, stock and rejected items in the feed, whether events are recorded without interruption and whether CPA moves within your threshold.

The permitted moves are narrow: switching off an ad set that is clearly dragging results down; raising the budget of a working ad set in steps; swapping a creative inside an existing ad set. The forbidden ones are those that move the structure or the way of bidding — a new campaign instead of a working one, another strategy, switching between CBO and ABO.

If something does have to be changed, you change one thing and wait before touching the next.

How do you read the result afterwards?

Compare with the same period last year, not with last month. A month-to-month comparison in a seasonal business is meaningless — it measures the season, not the work.

And look at whether the growth came from more new customers or from a larger average order. The two lead to different decisions for next year.

And the weak months?

They are the cheapest time for tests. The cost of impressions is low, competition is lower, and a mistake costs less. The creatives and structure you go into the peak period with are tested exactly then.

A test has rules so that it means something: one variable changed at a time, a period long enough to get out of the learning phase, and a written-down definition in advance of what you count as the result — CPA, ROAS or order rate, chosen before the start, not after it.

The weak months are also when the tasks forbidden at the peak get done: rewriting the tracking and reconciling it with real orders, cleaning up the feed and rejected items, changes to the checkout, refreshing audiences and lookalike lists, trying new creative formats.

Do not stop everything, though. Keep a base campaign with a reduced ad budget so you do not wipe out the accumulated learning, and direct the savings into tests. Record the result of every test, so you enter the peak period with ready answers, not hypotheses.

FAQ — Frequently asked questions about seasonality

Should I stop advertising in the weak months?

A complete stop wipes out the accumulated learning and puts the campaign back at zero. A reduced budget focused on tests is better.

How much more expensive is it at the peak?

It depends entirely on the industry and the year. What counts is your own comparison with the same period last year, not outside averages.

Does the seasonal bid adjustment work?

For short, predictable events — yes. For the whole season — no; it is a tool for a few days, not for a month.

When should I launch the promotion?

Earlier than the competition, if the margin allows. Switching on late means entering at the most expensive moment of the auction.

How do I plan for a new business with no history?

From search data and the industry calendar, but with smaller steps and more frequent monitoring — the first year is measurement, not planning.

Conclusion

Seasonality is not managed in the season. The work that decides the peak period is done in the weak one — tests, prepared creatives, a clean feed and tracking you can trust.

If you are heading into a peak period and are not sure the foundation will hold, start with an independent review.

Request a quote for a digital audit →

Related articles: campaign scaling · disapproved items · ROAS threshold