Q1 2026 Prep: January Auctions and the New Attribution Defaults

The holidays are not over, but your January is already being decided. What you export, rename and write down this month determines whether Q1 starts on clean numbers or on confusion.

Two things line up in the first half of January. The auctions often get calmer after the shopping peak. And on January 12, 2026, Meta says its Ads Insights API will stop returning the 7-day view and 28-day view attribution windows.¹ A plan for both, built now, turns a messy start into the best-run quarter you have had.

Why does January matter so much for Meta advertisers?

Because it is the first clean read of the year. Holiday promotions, shipping deadlines and gift buying distort every number from October through December. In January the noise drops and you can see what your business does on an ordinary day. (The pattern is covered in The Q1 CPM Dip.)

Demand this season was strong. Adobe measured $14.25 billion in US online spending on Cyber Monday 2025, up 7.1%, and $44.2 billion across Cyber Week, up 7.7%. Mobile drove 57.5% of Cyber Monday sales, and Adobe reported AI-driven traffic to US retail sites up 670% on Cyber Monday, a number worth watching.³

Meta, for its part, guided to fourth quarter revenue of $56 billion to $59 billion, which tells you how much advertising money was chasing the same holiday attention.⁵

Will January auctions be cheaper?

Often, in my experience, because a lot of advertisers pull back after the shopping peak. But not always, and not for every category. Fitness, finance and health offers can be busier in January because of new-year resolutions. The only honest answer comes from your account: open last January next to last December and look at CPMs, click-through rate and cost per purchase. Plan from that.

What changes on January 12?

Meta published the details on its developer blog on October 16, 2025.¹ The long version is in Meta Is Removing View Attribution Windows.

• Removed from the API: the 7-day view and 28-day view attribution windows.
• Still available: 1-day click, 7-day click, 28-day click, 1-day engaged view and 1-day view.
• New history limits: 13 months for unique-count and hourly breakdowns, 6 months for the frequency breakdown, and asynchronous-only for the marketing mix modeling breakdown.
• Unchanged: total values remain available for up to 37 months.

Meta describes the change as consistent with Ads Manager behavior, so if you only read Ads Manager, you will see little difference.¹ The risk sits with anything that pulls data through the API. PPC Land reports that requests for removed windows return empty data rather than errors, so a connector can go quiet without a warning.² Meta's post does not say what the response will be, so test your own tools.

What should you do before the end of December?

Treat the next four weeks like a reporting migration, because that is what it is.

Export before the cutoff, not after it. Once the limits apply, the older breakdown history may no longer be retrievable through the API.

What if a report goes quiet on January 13?

Run a three-step diagnosis before you touch a single ad. First, pull the same dates from Ads Manager and from the report. If Ads Manager shows normal results and the report does not, the problem is the report. Second, check which windows and breakdowns the report requests. If it asks for a removed window, replace it. Third, annotate the date in your log so nobody reads the dip as a performance problem in February. A quiet report is a reporting question first.

How should you reset budget and creative for January?

Use the calm to test, not to coast.

Reset budget in steps

After a holiday push, budgets are usually high and audiences are saturated. Step down gradually rather than all at once, and look at frequency and cost per purchase as you go. Then rebuild toward a budget your normal weeks can support.

Launch new concepts

Q1 is the best time to try ideas you did not have room for in November. Write three genuinely different angles, ship them together, and judge them over a fair window. Genuinely different beats recolored (see Meta's GEM Model): Meta's engineers described a model they call GEM in November as passing what it learns to hundreds of user-facing models.⁶ My read, not Meta's claim: distinct ideas give a system like that more to match to different people.

Jon Loomer added a useful brake in a video posted on December 3, 2025, aimed at people who read "more creative" as "more ads":

He is not against variety. He is against drowning yourself in volume. Three real concepts beat thirty rushed ones.

Plays here, from YouTube

Let the offer reset

Many people are making fresh starts in January. If your offer fits that mood, say so plainly. If it does not, resist forcing a new-year angle.

How should you set Q1 targets on clean numbers?

Start from facts, not from December.

• Use a benchmark week, not a holiday week.
• Set targets in cash collected, reconciled against your store, processor or CRM.
• Hold the definition steady for the whole quarter. Do not change windows mid-quarter.
• Annotate changes so a dip in the data can be traced to a decision.

Treat incremental attribution as a side lens. Seer Interactive's test of Meta's setting across about $1.05 million of spend found Meta classified 87% of conversions as incremental, but Seer advised cross-checking with independent sources.⁴ I would run it next to your standard reporting for a quarter before trusting it (Incremental Attribution explains why).

What does the first month look like?

• Before January 1: exports done, reports rebuilt, benchmark saved.
• January 1 to 11: step budgets down, launch new concepts, annotate everything.
• January 12: note the date in your log. Check that your reports still return what you expect.
• January 13 to 20: compare reports to Ads Manager for the same dates. Fix any gap.
• January 21 to 31: review concepts, double down on the best, cut the rest.

If you do the quiet work this month, January will not be a scramble. It will be the first week you can read your business without static. That clean read is what the systems I build for clients are designed to give.

Keep building.Ekai