Paid social for B2B, and the narrow case where it works

The pitch for paid social to a B2B company always starts with the targeting. You can put an ad in front of VP of Engineering at companies with 200 to 2,000 employees using Kubernetes, and nobody else. That precision is real, and it is the entire reason the conversation happens at all — nothing else in advertising lets you specify a stranger by job title and tech stack before you have spent a cent. The people selling the channel are not wrong about that part.

What they leave out is that precision answers only one of the two questions a channel has to answer. It tells you who saw the ad. It says nothing about whether the moment they saw it was one where being sold to made sense. A person scrolling LinkedIn between meetings is not looking for a customs broker, a data warehouse or a fractional CFO. They are looking at what their old colleague posted. Reaching the right VP of Engineering at the wrong moment is not a lesser version of the thing search advertising does — it is a structurally different mechanism, and it converts at a different order of magnitude because of it.

That is the honest starting point, and it is worth sitting with the strongest form of the counter-argument before making the case against the channel, because the counter-argument is correct as far as it goes: interruption advertising built brand recognition for a century before anyone could target a search query, and plenty of large B2B companies run paid social profitably today. The channel is not broken. It is narrow, and most companies that try it are outside the narrow part.

The number that matters is not the one on the dashboard

Every paid social platform reports cost per lead, or cost per click, or cost per completed form. None of those numbers is the one that determines whether the spend made money. The number that matters is cost per qualified opportunity — a lead that a salesperson looked at, judged real, and put into a pipeline with a plausible reason to buy.

The gap between the two is usually where the arithmetic falls apart. Say a campaign produces leads at $80 each, which sounds reasonable next to the CAC benchmarks in most decks. If one lead in eight from a cold interruption channel turns out to be a real buyer — a generous rate for someone who did not search for anything, was not referred, and clicked because a headline caught them mid-scroll — the cost per qualified opportunity is $640. Run that number against your close rate and your average contract value, not against the $80 figure the ad platform is proud of. A company with a $3,000 average deal and a one-in-four close rate on qualified opportunities needs four opportunities to land one customer, at $640 each: $2,560 to acquire a customer worth $3,000. That is not a channel. That is a company paying to lose money slowly, dressed up as marketing.

The same arithmetic run against a $40,000 average contract value looks completely different. Four opportunities at $640 is $2,560 against a deal worth sixteen times that. The targeting did not get any more precise between the two examples. The only thing that changed is what the company was selling, and that single variable is doing almost all the work in whether the channel is viable.

The contract value that makes it survivable

There is no fixed number that separates "paid social works" from "paid social doesn't," because it depends on close rate and lead-to-opportunity rate, both of which vary by company. But the shape of the constraint is fixed: the channel only pays when contract value is large enough to absorb the low conversion rate that comes from interrupting people who were not looking, rather than a sales cycle short and cheap enough that the low conversion rate does not matter.

Practically, that means paid social for B2B tends to work at two ends of a spectrum and fail in the wide middle. It works for enterprise software and services where a single contract is worth tens of thousands of dollars a year, because the channel only has to produce a handful of real opportunities a quarter to be worthwhile, and a handful is what interruption advertising is actually capable of. It also works, unusually, for very cheap self-serve products — a $20 monthly tool someone can start using without a conversation — because the offer matches the medium in a different way, covered below. What it reliably fails for is the company in between: a $2,000 to $15,000 annual contract, a sales cycle that needs a demo and a follow-up, sold to a buyer who was not looking when the ad appeared. That is most B2B companies, and it is exactly the range where the CAC arithmetic above stops closing.

What a stranger can accept, and what needs a warm list already

The offer matters as much as the targeting, and this is the part most campaigns get backwards. Paid social interrupts someone who was not in a buying mindset. Asking that person to book a call with sales is asking for the largest possible commitment at the worst possible moment in the relationship — the ad equivalent of a stranger proposing marriage on the first hello. It is not that the offer is bad. It is that the offer is mismatched to where the viewer's head actually is.

What works instead is something a stranger can accept on the spot with no ongoing commitment: a benchmark report with real numbers in it, a free calculator that does something useful in thirty seconds, a template that solves a small piece of the buyer's problem without requiring them to talk to anyone. These convert on cold traffic because the entire transaction completes in the moment of the click. "Book a demo," by contrast, is an offer that works beautifully on retargeting — someone who already visited your pricing page and is now being reminded — and works badly on prospecting, because you are asking a stranger for the commitment a warm lead gives after several touches. Companies that run the same call-to-action across both audiences are, in effect, running one good campaign and one bad one under a single reported number, and the good one is carrying the average.

Retargeting numbers flatter the whole account

This is the mechanism by which a paid social account can look profitable in the dashboard and lose money in the P&L. Retargeting — showing ads to people who already visited the site, opened an email, or sit on a customer list — converts far better than prospecting, for the obvious reason that the audience was already partway to a decision before the ad ever ran. That is not a criticism of retargeting; it is a genuinely efficient use of a small budget, and it is worth running on its own.

The distortion happens when retargeting and prospecting sit in the same campaign or the same reported blended CPA. A prospecting audience of cold strangers converting at a low rate, mixed with a small retargeting audience converting at a high rate, produces an average that looks respectable while hiding that almost none of the actual pipeline came from the part of the campaign that was supposed to be finding new customers. The company reads the blended number, concludes the channel works, and increases the prospecting budget — the exact part of the account that was not producing the result being credited to it. Separating the two budgets and the two reported numbers is not optional bookkeeping. It is the only way to know what you are actually paying for.

The platform counted a conversion that never reached your CRM

The last trap sits downstream of both of the others. LinkedIn and Meta report conversions inside their own dashboards using their own attribution logic — often a window of several days, often crediting a view of the ad rather than a click on it, and always counting an on-platform action like a form-fill rather than what happened to that person afterward. A form-fill is not a qualified opportunity. It is a data point that something might have happened.

The only number worth trusting is the one that comes out the other end: how many of those form-fills turned into a record in your own CRM that a salesperson could reach, and how many of those turned into a real conversation. Every company that has ever compared the two counts finds a gap, and the gap is usually large enough to change the CAC conclusion entirely — a campaign that looks like it produced sixty conversions at the platform level might have produced eleven contactable leads and two real opportunities once someone actually tried to reach the people on the list. Building that comparison once, by hand, for a single campaign is more informative than a quarter of dashboard-watching, and it is the same discipline this publication has argued for more generally in why attribution is mostly a story companies tell themselves.

The recommendation

Run paid social if your average contract value is large enough that a handful of real opportunities a quarter justifies the spend, or your offer is cheap and self-serve enough that a stranger can say yes without a conversation. Build the offer for the medium — something a cold viewer can accept on the spot — rather than porting over the call-to-action from your website, which was written for someone already convinced. Split retargeting and prospecting into separate budgets and separate numbers from day one, because a blended CPA will always tell you the channel works slightly better than it does. And before increasing spend on anything, trace one batch of platform-reported conversions all the way into your CRM by hand, because that is the only number immune to the platform's own incentive to look good.

Outside those conditions, the channel is not broken so much as being asked to do a job it cannot do — interrupt a stranger and walk them to a five-figure sales conversation in one motion. That gap is also where the other channel worth checking first tends to close faster: existing customers who already trust you are a warmer, cheaper route into the same buyer population, which we've set out in the mechanics of partnerships and referral fees that actually hold up. If your company already has a working referral motion, it is very often the better place to put the next dollar before it goes anywhere near an ad account — a point this publication has made about acquisition channels more broadly in looking at which channels survive scrutiny and which only survive the meeting where they were proposed.