The campaign didn't change. The calendar did. I just wasn't watching it.
The setup
— Gifting/e-commerce offer, $7 payout, native traffic, scaled hard in November.
— Peak: ~$3,000/day spend at 60% ROI. Roughly $11k profit for the month.
The move
It was working so well I let it ride into January on autopilot, same budget, same creatives.
The numbers (illustrative)
— Mid-December: ROI still +45%.
— Jan 2-15: ROI -18%. Buyer intent evaporated post-holiday, CPMs stayed high from the Q4 hangover.
— ~$3k of December's profit handed back in two weeks.
The lesson
Seasonal campaigns have an expiry date the dashboard won't show you. Autopilot is fine on a flat trend and lethal across a demand cliff. The same ad that printed in November was selling sleds in spring.
What I'd do differently
Set a hard calendar-based wind-down on seasonal offers, decided in advance, not a "watch the ROI" plan. By the time the ROI line drops, you've already paid the post-season CPM premium for a week.
Arb Files
@ArbFiles
The Q4 offer that made $11k in November and -$3k in January
This post was published in the Telegram channel Arb Files (@ArbFiles). All its posts are on the channel page.
Open on Telegram