Peak Season: What To Decide Before The Auction Gets Expensive

The short version

During peak season, costs rise and everyone competes for the same attention. The decisions that determine your outcome are mostly made before it starts, because once CPMs climb your options narrow to bidding — the least powerful lever available.

The things worth doing early: build audiences, prove creative, exit the learning phase, and decide your margin floor while you can still think clearly.

Before: the work that actually matters

Build retargeting pools early. The people you will sell to at peak should already know you. Prospecting during peak costs the most it will cost all year; prospecting before it costs the least. This is the single highest-return pre-peak activity.

Prove creative in advance. Testing during peak means paying peak CPMs to learn things you could have learned cheaply. Arrive with winners already identified.

Exit the learning phase before it starts. A campaign entering peak in learning spends the expensive period exploring. Get structure and optimization events settled early, then stop editing.

Fix the feed and the site. Availability accuracy, page speed, checkout. Peak traffic exposes every weakness at the worst moment.

Decide the margin floor. Write down the point at which you stop bidding. Deciding this in advance, calmly, is the difference between a disciplined peak and one where the ROAS target drifts downward daily because volume feels urgent.

During: what to actually do

Expect worse reported efficiency. Higher CPMs mean higher CPA at the same performance. Judge against your margin floor, not against last month.

Do not restructure. Any meaningful change resets learning at the exact moment you cannot afford exploration.

Monitor pacing daily. Budgets exhaust faster than usual. Front-loading that was harmless in October costs real revenue in December.

Watch inventory alongside spend. Advertising products you cannot ship is the most avoidable peak mistake.

Retarget hard. Your existing audience is the cheapest thing you own during the most expensive period of the year.

After: the part everyone skips

Do not cut spend to zero immediately. Post-peak has cheap inventory and buyers with gift cards and returns to exchange. It is often the most efficient window of the quarter, and it is usually abandoned because budgets were annual.

Watch returns. They arrive weeks later and will reduce the peak ROAS you already reported and celebrated.

Segment the cohort. Peak buyers often behave differently — many are gift purchasers who never buy again. Applying normal repeat assumptions to them overstates what peak was worth.

Write down what happened. Which creative worked, which price points moved, what CPM you actually paid. Next year this is the plan.

Practical rule

Do the work in the cheap weeks. Once CPMs rise, you are executing a plan rather than making one, and the only real decision left is whether to keep bidding — which is why the margin floor should be written down before you need it.