Meta says ad sets under 50 conversions a week cost 28% more per purchase. B2B does not produce 50 of anything in a week.

Meta put a number on it. “Give each ad set enough volume to learn from. Roughly fifty conversions a week,” the company wrote in an October 21 post reported by PPC Land, adding that “in our analysis, ad sets under that threshold ran about 28% more expensive per purchase.” It shipped the arithmetic too: divide the weekly budget by the cost per result multiplied by fifty, and that is how many ad sets it can carry.

For a retailer moving $40 sneakers, that is housekeeping. For B2B it is a price list.

The threshold is old news. Roughly fifty optimization events in seven days has long been the documented condition for an ad set leaving the learning phase. What is new is Meta attaching a cost to missing it, in public, in the weeks when next year’s budgets get drafted. And no amount of trying harder gets B2B to fifty. The threshold describes a different business.

Meta will not say how it got to 28%

Meta published no methodology with the figure. No sample size, no time period, no region, no vertical, no campaign type, and nothing on whether budget was held constant. That last omission matters most: an ad set with low conversion volume is usually an ad set with a small budget, and a small budget explains both the thin volume and the higher cost. Meta has described a correlation and priced it like a mechanism. We have been here before with this company, whose shopping agent decides which brands a buyer sees and publishes no ranking rules either.

Run the formula at B2B prices and one ad set wants $15,900 a week

First Page Sage’s October 8 report puts paid cost per lead at $318 in B2B SaaS and $518 across the technology sector, from lead data collected between January 2022 and August 2026. Two caveats first: the firm does not disclose how many companies sit behind those averages, and the figure is blended across all paid channels rather than Meta alone. Read it as what a B2B lead costs these companies, not what Meta charges for one. It is also the number a CFO has in mind. At $318, fifty conversions a week costs $15,900. For one ad set. By our arithmetic on Meta’s formula, that is $826,800 a year to keep one ad set out of the learning phase, before anyone asks whether the leads were any good.

Meta’s own ground is the fairer comparison and does not rescue the picture. WordStream’s 2026 Facebook benchmarks, built from 452 US lead campaigns between April 2025 and June 2026, put the median cost per lead in industrial and commercial at $35.87. Fifty of those a week is $1,793, roughly $93,000 a year for one ad set. That is the floor, not the bill, because a Meta form fill is not a qualified opportunity.

B2B raised its Meta share to 24.9% anyway

Primer’s State of B2B Paid Advertising 2026, built from observed ad accounts at 379 B2B advertisers over the twelve months to August 2026, found Meta’s average share of paid budget rising from 20.8% to 24.9%, with adoption up from 54% to 63%. Of that Meta spend, 89% goes to the lead and sales objectives the threshold governs. Primer sells targeting and measurement software and says its panel is not a representative census, the right caveat to carry.

The same report shows how thin the ground is. In Primer’s incrementality tests the comparison groups, the people not shown the ads, produced fewer than one form submission per 100,000. That baseline sits so close to zero that 92 of the 140 advertisers with incrementality evidence could not produce a defined lift result in ninety days. Meanwhile platform-reported conversions exceeded confirmed website conversions several times over on Meta.

Set that beside the threshold. The conversions counted toward your fifty are not the ones you can confirm.

From outside the industry it is a strange arrangement. The platform sells a machine that learns, then bills you for the lessons it could not finish.

The channels that charge nothing for low volume

None of this means B2B is bad at paid social. The shape is wrong. A buying committee at three hundred accounts is not fifty conversions a week, and the 28% is what the system charges for that fact. The same pattern showed up in prospecting, where sales leaders rank AI-written email dead last of seven channels at 8% while cold calling holds 47%. The channel that scales cheapest is rarely the one that works.

Which leaves the uncomfortable comparison. The one channel that closes deals is the only one with no reporting attached, and the only one where twelve of the right people in a room counts as a result rather than a failed learning phase. Event budgets moved that way this year, with spend per attendee up 28% and rooms getting smaller. The work left there is instrumentation, not volume, and exhibitions already absorb 40.8% of the average exhibitor’s marketing budget while staying the hardest line to defend.

The position. The teams that cut paid social to one deliberately underfed retargeting ad set this quarter and move the difference into rooms they control will show a better cost per opportunity by Q2. The teams that chase fifty will spend 2027 funding the learning phase and calling it demand generation.

Undivided Attention is powered by Mobilo.