How a seasonality adjustment works
Smart Bidding sets bids from an expected conversion rate, learned from your account’s recent history. A seasonality adjustment overrides that expectation for a window you define. You choose the campaigns, the start and end times, and how much higher or lower you expect conversion rate to run, and bidding behaves accordingly for that period before returning to normal on its own.
It applies to the automated strategies — Target CPA, Target ROAS and the Maximize options — and it is created in the Google Ads shared library rather than inside a campaign. The window is deliberately short. Google built it for events measured in days, and the interface will not let you stretch one across a whole season.
Why seasonality adjustments matter
Smart Bidding is reactive. It notices that conversion rate has jumped only after conversions start arriving, which means the opening hours of a short sale get bid as though it were an ordinary day. For an event that lasts a weekend, that lag can swallow most of the event.
The adjustment closes the gap by telling the system something you already know and it cannot. This is one of the few places left inside an automated bidding strategy where your knowledge of the business genuinely beats the model, because no model has any way of knowing about tomorrow’s offer.
Where seasonality adjustments go wrong
The most common misuse is treating it as a growth lever. It adds no budget and creates no demand; it only changes what a click is assumed to be worth. On a campaign already limited by budget it changes very little, and reaching for it when the real constraint is spend is a way of looking busy rather than fixing anything.
The second is applying it to long periods. A festival season in Nepal runs for weeks rather than days, and across that span Smart Bidding will learn the real pattern from real conversions more accurately than anyone can estimate it in advance. The third is guessing the size of the change. An estimate set too high bids aggressively for traffic that was never going to convert, and the cost of that shows up within days.
How to use it properly
Use it only when three things are true: the event is short, it is planned, and you can point to evidence for the size of the change. Look at what your own account did during the same event last year and let that set the estimate, rather than optimism about this year.
Set the end time at the same moment you set the start, so a temporary instruction cannot outlive the event that justified it. Afterwards, compare the period against a normal one and note whether the estimate was close, because that record is what makes the next one better. Keep it clearly separate from a data exclusion, which exists to tell Smart Bidding to ignore a stretch when conversion tracking was broken. If you find yourself reaching for either control often, the account usually needs steadier campaign management rather than more manual overrides.
