Meta published a piece of arithmetic last Tuesday that most paid social accounts will fail. It is one line long, it uses numbers you already have, and it tells you how many ad sets your budget can actually feed. Run it with November prices instead of September prices and the answer gets smaller, which is the whole problem.
The formula sits inside a Meta for Business post dated October 21, titled “Peak Season 2026: You Don’t Beat The Auction. You brief it.” Most of the post is an essay about treating campaign settings as a brief. The part that should change what you do this week is three sentences near the end.
One line of math
“Give each ad set enough volume to learn from. Roughly fifty conversions a week. Below that the system is guessing,” Meta writes. “In our analysis, ad sets under that threshold ran about 28% more expensive per purchase.” Then the formula, presented in bold:
Total weekly budget ÷ (cost per result × 50) = the most ad sets your budget can support
Meta follows it with a line that reads like a diagnosis: “Most accounts are running more than that. Which means most accounts have written a brief with more objectives than they have money to pursue.”
The fifty is not new. Meta’s own help page on the learning phase says ad sets usually exit learning “after about 50 results in the week after the ad set’s last significant edit,” and that during learning they “usually have a higher CPA.” The same page tells advertisers to avoid high ad volumes and to combine similar ad sets, because “by combining similar ad sets, you also combine learnings.” What is new is that Meta has turned an old rule of thumb into a hard ceiling with a price tag attached, and published it five weeks before Black Friday.
Here is how fast the ceiling moves. PPC Land worked the example: a $20,000 weekly budget at $40 per purchase needs $2,000 per ad set per week to buy fifty conversions, so it supports ten ad sets. At $100 per purchase, the same budget supports four. Extend that yourself: if your purchase cost drifts from $40 to $50 in November, ten becomes eight. Nobody deleted two ad sets. The math did.
Why November makes the ceiling drop
The formula has two inputs that move in opposite directions in Q4. Budget usually goes up. Cost per result usually goes up too, and if it rises faster than budget, you can afford fewer ad sets in the month you are trying to do the most.
The auction is about to get crowded, and not only with your category rivals. Digiday’s Sam Bradley reported on Monday that political clients spent $149 million on YouTube and Meta in the week beginning October 11, up from $34 million in the final week of July, citing data firm AdsMom. Google took $97 million of that, which leaves roughly $52 million on Meta in a single week. One agency executive told Digiday they had seen average YouTube CPMs rise about 20% in recent weeks. Election day is November 3. Holiday money keeps coming after it.
Meanwhile brands are moving budget toward exactly the place you buy. In the same day’s Media Buying Briefing, Sarah Engel, president of January Digital, put it plainly: “In Q4, if it can’t be measured, it isn’t making the plan.” She said that attitude is driving spend toward Meta, TikTok and Amazon. Ashley Muench of Scène Studios said her DTC and apparel clients’ budgets were up as much as 20% on last year, mostly headed to social. WARC’s forecast, cited in the same piece, has social spend growing 21.3% in Q3.
More money chasing the same feeds is good for Meta and neutral at best for your cost per purchase.
The trap in the denominator
The moment you run the formula, a shortcut appears. If purchases cost $100 and you need fifty a week per ad set, why not optimize for add to cart at $15 instead? Suddenly every ad set clears the bar and the dashboard stops saying “Learning limited.”
Do not do this, and Meta’s own post explains why. Most accounts, it says, “optimise for what’s easy to count, not for what the business gets paid on.” Swapping a purchase event for a cheaper event to satisfy a learning threshold is the purest version of that mistake. You will get fifty of something every week. It will not be the thing your finance team counts.
The honest fix is fewer, bigger ad sets on the event you actually want. If your account cannot afford even one or two ad sets at fifty purchases a week, that is not a structure problem. It is a budget conversation, and better to have it on October 26 than December 2.
One exception worth knowing: Meta’s help page says Shops ads need “a minimum of 17 purchases through your website and 5 through Meta” to complete learning. If you sell through Shops, run the math against that bar, not the round fifty.
What the 28% is, and what it is not
Be clear about the evidence before you rebuild an account around it. Meta gives no sample size, time period, vertical or method for the 28% figure. PPC Land points out the obvious confounder: ad sets with small budgets are both less likely to hit fifty conversions and more likely to have different cost structures for reasons that have nothing to do with learning. The same piece notes vendor reports, without Meta documentation, of lower thresholds for some campaign types.
There is also a commercial reading. Fewer, larger ad sets with broad settings is the shape of account Meta’s automation prefers, and it has been pushing advertisers in that direction for years. A post that tells you to consolidate is consistent with Meta’s interest whether or not the 28% holds for your account.
None of that makes the formula wrong. It makes it a planning tool rather than a law. The arithmetic of spreading a fixed budget across too many ad sets is not controversial. The size of the penalty is Meta’s claim, and the only way to know it for your account is to measure it yourself, which, to its credit, is exactly what the post asks you to do.
Consolidation is an edit, so the clock matters
Here is the part that turns a nice idea into a deadline. Merging ad sets means building new ones or making significant edits to old ones, and Meta’s help page says both put an ad set into learning. Meta’s help page says ad sets exit learning after about fifty results “in the week after the ad set’s last significant edit.” So every structural change you make costs roughly a week of less stable delivery and higher cost per acquisition.
Meta’s post says you pay for every change twice: once for the learning phase you restarted, and once for the old instruction, which was “live and spending at full price the whole time you had it wrong.” Its advice is to “group your changes. If you’re going to rebrief, do it all at once. Take one learning phase instead of five.”
Black Friday is November 27 this year. If you consolidate this week, your new ad sets have about four weeks to settle before the most expensive days of the year. If you consolidate the week before Thanksgiving because the dashboard finally scared you, you will be learning at full price on the days that matter most. The calm window is now.
The post also asks you to hold your biggest campaign still and measure it with “an A/B test or a Conversion Lift study,” and to “give it the month of October.” It was published on October 21. You do not have the month of October. You do have most of November, and a held-out test that runs from early November through Cyber Monday will tell you more about your account than any figure in Meta’s post.
The rebrief you did not write
Meta’s essay describes the auction as something “rebriefed by every edit you make. Including all the ones you don’t think of as edits.” That applies to Meta’s own changes too, and one lands this Wednesday.
Starting October 28, Meta will add product browsing to ads that are already running on Facebook and Instagram unless advertisers opt out, according to a notice PPC Land reported on October 15. “When people tap your ad, they may see an experience on Facebook or Instagram with product info and personalized suggestions,” the notice reads. Advertisers can opt out through a button in the notice before October 28, or turn it on or off “anytime in your Advantage+ Creative and Destination settings.” Meta backs the change with a claim of 3% more conversions from an August test of about 117,700 advertisers, without saying which conversions or at what cost.
Whether you keep it or not is your call. The point is that it changes what happens after the tap on ads you may have just consolidated, in the same week. Media buyer David Herrmann, whose LinkedIn post made the notice public, said a client’s traffic on an ad they thought was working turned out to be a collection ad on Instagram Reels, not the creative. His ask to Meta was to “create a column or a notice that product browsing is on and contributing X to the ad.” Today there is no such column. If you are holding a campaign still to measure it, a platform default flipping underneath it is exactly the kind of edit that muddies the read. Decide on purpose, write down what you decided, and keep the setting the same through your test window.
What to do this week
Run the formula on every account today, with November numbers. Take next week’s planned weekly budget and a realistic November cost per purchase, not the October average. Divide. Write the result next to the number of active ad sets you are actually running. If the second number is bigger, you have your list.
Keep the event, cut the ad sets. Do not switch optimization to a cheaper event to clear the bar. Merge similar ad sets until each one can plausibly reach fifty of the outcome you care about. If even that does not work, take the gap to whoever owns the budget this week, with the arithmetic attached.
Do every structural change in one batch, before November 1. Consolidation, optimization event changes, attribution setting changes, bid strategy changes: all at once, one learning phase. Then freeze structural edits and let creative rotation be the only thing that moves through Cyber Monday.
Decide product browsing by Tuesday night. Check every account the notice might cover, including client accounts if you are an agency. Opt out, keep it, or keep it on a subset, but make the choice in writing so nobody misreads a change in results next month.
Set up one held-out test on your largest campaign. An A/B test or a Conversion Lift study, started in the first days of November, held still through Cyber Monday. That gives you your own answer to the 28% question instead of Meta’s.
Write the three-line brief. Meta’s suggestion is a good one: for each campaign, write down what it optimizes for, how hard it should chase, and how results are counted. If the attribution setting does not match the number finance uses, fix that now, not in the January post-mortem.
Meta’s post ends by saying peak season rewards “the one who wrote the clearest instruction and then had the nerve to leave it alone.” That is true. It is also easier when you have done the division first and know how many instructions your budget can afford.
