The Q1 CPM Dip: How to Use January's Cheaper Auctions

Meta's revenue fell about 11% from the fourth quarter of 2022 to the first quarter of 2023. It fell about 9% from the fourth quarter of 2023 to the first quarter of 2024.

Same company, same seasonal step down, two years running. (If you read my Black Friday pacing plan, this is the other side of that peak.) I do not think that is a coincidence, and I think service businesses with a ready offer can use it. I also want to be careful about how far the evidence goes, because there is a lot of confident nonsense about ad costs this time of year.

The January discount is a pattern, not a promise. Here is what is solid, what is inference and how to be ready on January 2.

Is there really a January dip in Meta ad costs?

There is a clear dip in spending. The dip in price is an inference.

What the filings show

Meta reported Q4 2022 revenue of $32.17 billion and Q1 2023 revenue of $28.65 billion, a drop of about 11%.¹² It reported Q4 2023 revenue of $40.11 billion and Q1 2024 revenue of $36.46 billion, a drop of about 9%.³⁴

In both cases Meta's own Q1 outlook, given when it reported Q4, was below the Q4 number: $26 to $28.5 billion in the first case and $34.5 to $37 billion in the second.¹³

What they do not show

Revenue is price multiplied by volume. Meta does not break out a monthly price curve, and its reported change in average price per ad is year over year, not quarter to quarter. So I cannot hand you a sourced January CPM number, and I would be wary of anyone who does without showing their data. What the revenue pattern supports is narrower: a big chunk of advertiser demand leaves after the holidays.

Why might auctions get cheaper in January?

Because there are fewer bidders chasing the same attention.

Demand leaves first

Late in the year, retailers push hard. Meta guided Q4 2024 revenue to $45 to $48 billion, compared with $40.59 billion in the third quarter.⁵ I broke that quarter down in Q3 2024 and the election squeeze. When that pressure ends, the same number of people are scrolling and fewer advertisers are competing for them. All else equal, that eases the price of reaching those people.

All else is not equal

I say "all else equal" on purpose. Audiences, competition in your niche, your creative and your offer all matter more to your cost per client than the calendar does. The calendar changes the weather. It does not change the quality of your boat.

What do you need ready on January 2?

Everything that turns cheaper attention into booked calls.

How much budget does January need?

Enough to get out of the learning phase, or the discount is moot.

The learning phase rule

Per Meta's Help Center, an ad set leaves the learning phase once delivery is stable, which typically takes around 50 optimization events within a week of the last significant edit. It also says that during learning, performance is less stable and cost per result is usually higher.⁶

A worked example

These numbers are made up. Say you expect each booked call to cost $40. Fifty results a week is $2,000 a week, about $286 a day (the same logic as my creative testing kill rule). If your plan allows only $100 a day, the ad set will likely sit in learning and look worse than it is. In that case, optimize for an event that happens more often, or run fewer ad sets so the budget is concentrated.

The same Help Center page advises against frequent budget changes and unnecessary edits, because either can restart learning.⁶ Set the plan on January 1 and let it run for the first week.

How do you test the dip in your own account?

With a plain before-and-after, written down.

Three numbers, three windows

Pull your cost per thousand impressions, your cost per booked call and your show rate for three windows: the last two weeks of November, the first two weeks of December, and the first two weeks of January. Put them in a small table. Do not look at one number alone. A cheaper CPM that comes with a worse show rate is not a discount.

Keep a note for next year

Write down the date you launched, what changed in the account and what you saw. Memory is a poor record of ad costs. A note beats a feeling, and in a year you will have your own seasonal curve instead of someone else's guess.

What can go wrong?

Cheap attention does not equal cheap clients.

Soft conversion

The first days of January can bring curious clicks and slower decisions. Do not scale on day two. Give each campaign at least a week and judge it by booked calls, show rate and collected cash.

Poor offers get no help from a discount

A weak offer in a cheap auction is still a weak offer. If your January page reads like your December page, change it. Make the start date and the reason to move now concrete.

I plan every year around the same idea: buy attention when it is cheap only if you are ready to turn it into clients. If you are, January is a good month to be a prepared operator. If you want a second opinion on your January plan, reach out.

Keep building.Ekai