Picture the mistake that plays out in some ecommerce account every Black Friday: a manager pulls a big chunk of budget off a Meta campaign six hours into the day because the dashboard shows ROAS has cratered.
The campaign almost certainly hasn't cratered. The conversions behind that ROAS number haven't finished arriving yet. If the real number would have settled fine a day later, the campaign is now underfunded for the back half of the day it mattered most, and whoever pulled the budget has to re-add it mid-Cyber-Monday under worse auction pressure than it left under.
That is the actual failure mode behind most bad Black Friday ad budget pacing decisions. Not the wrong day-by-day split. Not a Google Ads vs. Meta allocation mistake. Reacting to a number mid-sale that had not finished settling, and making a budget call off it anyway.
The standard budget-split playbook isn't the problem
Most Black Friday advice reduces to a shape most paid media teams already roughly follow. One illustrative version: ramp spend starting 10 to 14 days out so campaigns exit the learning phase before the peak, then weight the five-day Cyber Week window itself toward the two biggest days, something like 35 percent of the week's budget on Black Friday and 30 percent on Cyber Monday, with the remaining 35 percent spread across Thanksgiving and the weekend between them. Prospecting usually gets the larger share early in the week, then cedes ground to retargeting as auction pressure and CPMs climb toward Cyber Monday. Treat those splits as a starting hypothesis to test against your own account's prior-year data, not a formula.
2025 Cyber Week data puts the scale that shape is built for in context: Black Friday alone drove $11.8 billion in online spend, up 9.1 percent year over year, and Cyber Monday broke its own record at $14.25 billion, the single biggest online shopping day on record. A pacing plan built for that kind of volume surge is not wrong to exist.
What the standard playbook skips entirely is the measurement problem sitting underneath every pacing decision made while the sale is live. The plan tells you how much to spend on Friday versus Saturday. It says nothing about whether the number you are using to judge Friday's spend is even finished arriving by the time you look at it.
What's actually safe to react to mid-sale, and what isn't
Not every metric on a Black Friday dashboard updates on the same clock, and treating them as if they do is where pacing goes wrong.
Spend and impressions are close to real time. If a campaign is pacing 30 percent ahead of its daily budget by noon, that number is accurate right now. Spend-to-plan is one of the few signals on the dashboard you can trust at face value during the event itself.
Conversion-based metrics are not. GA4 takes 24 to 48 hours to fully settle on a normal day, and Meta's own help documentation confirms its event counts can differ across reporting surfaces before everything finishes processing. The point mid-sale is narrower than the full mechanic: the ROAS and cost-per-purchase figures on the dashboard four hours into Black Friday are built from a transaction count that is still climbing, not a finished total, and that gap is widest exactly when conversion volume is spiking fastest.
A full reconciliation against the store's own order export, the way our post-BFCM audit checklist covers, is the right move once the weekend is over. Mid-sale, the goal is smaller: know which numbers on the dashboard you can act on right now, and which ones you have to wait on.
A pacing framework built around when the data actually settles
Run pacing decisions against checkpoints, not against whatever the dashboard shows at the moment someone is looking.
- Pre-event (T-14 to T-3 days): Confirm tracking is frozen and correct before volume hits it. This is what our BFCM tracking checklist covers, including the duplicate-purchase-event fix that matters far more at 10x normal traffic than it does in a slow week.
- Hour 0 to 24 of the event: React to spend-to-plan and impression share only. Do not reallocate budget off ROAS or cost-per-purchase in this window. If spend-to-plan shows a campaign badly under-pacing, that is a delivery problem (budget, bid cap, or audience size), not a performance verdict, and it is safe to raise the budget or loosen a bid cap to fix delivery without judging performance yet.
- 24 to 72 hours in: Treat conversion metrics as directional, not final. A campaign that still looks weak after a full settlement window has passed is a more credible signal than one that looked weak four hours in.
- Past 72 hours: This is the earliest point conversion-based metrics are stable enough to make a real budget verdict on, consistent with how long Meta's modeled conversions take to finish arriving. Reallocating budget between campaigns based on ROAS before this point is reacting to noise that happens to look like a trend.
Resist the urge to make several small live edits to a struggling ad set during this window too. Meta treats a large enough budget change as a significant edit, and stacking a handful of "small" panic edits across Black Friday weekend can reset an ad set's learning phase repeatedly, which is its own cost separate from the pacing decision itself. Our piece on what actually resets Meta's learning phase goes into exactly where that magnitude line sits and what counts as significant, which is worth knowing before touching a live ad set mid-sale.
Let Smart Bidding handle the seasonality, not a manual panic edit
Google Ads has a built-in mechanism for exactly this kind of short, high-intent event: a seasonality adjustment. It tells Smart Bidding to expect a temporary change in conversion rate over a defined window, which is a better fit for a one-to-seven-day event like Black Friday weekend than an extended seasonality period.
Google's own guidance is specific about calibration: set the adjustment based on the expected change in conversion rate, not spend, and avoid leaving it running longer than the event actually lasts. A seasonality adjustment left active for two weeks after Cyber Monday ends is still telling Smart Bidding to expect a conversion spike that is no longer happening.
The point of using it is the same point behind the checkpoint framework above: give the bidding system a documented, deliberate instruction ahead of the event, instead of making a series of manual bid or budget edits mid-sale in response to numbers that have not settled.
The one number to watch before you touch a live budget
If there is a single rule worth carrying into Black Friday weekend, it is this: spend-to-plan is the only number on the dashboard safe to act on in real time. Everything downstream of a conversion, ROAS, cost per purchase, revenue per campaign, needs its settlement window before it means what it looks like it means.
That does not mean waiting passively. It means fixing delivery problems immediately when spend-to-plan flags them, and holding every reallocation decision until the conversion data behind it has actually finished arriving.
For teams who want someone watching pacing and edit timing through the weekend itself rather than reconstructing it after the fact, that is what our ads management service is built to do. If Black Friday pacing problems keep tracing back to data you couldn't trust in the moment, a paid media audit checks tracking and attribution health before the next peak period puts it under the same pressure again.
