There is one number that always makes a B2B advertiser smile: on Meta, a lead costs next to nothing. On the campaigns I ran for DecisionIA, an AI training provider aimed at company directors, the Meta channel came out at a cost per acquisition of €3.73 where Google started at €7.71. Half the price, same target, same objective. The temptation is immediate: shift the budget and scale.
That is exactly where the trap springs. The number is true, and it means nothing until you hold it against what the lead becomes. Because Meta does not target job titles: it targets people. Where LinkedIn Ads does the qualifying inside the targeting, and charges a premium for it, Meta sells volume cheaply and leaves the sorting to you. The only question is where that sorting will happen. Because it will: either upstream, in the creative and the form, or downstream, in your sales team's time.
Meta does not sell cheap or expensive leads: it sells the volume you ask for, at the cost you accept. The only number that commits to anything is the cost of the lead that ends up buying.
The cheap lead is Meta's default product
You need to understand why the cost per lead is so low, mechanically. When you optimise for the “form submitted” event, the algorithm does exactly what it is told: it looks for the people most likely to submit a form. Not to buy, not to show up for a meeting: to submit a form. And it is very good at that game.
The format helps. Meta's instant forms pre-fill the fields from the profile: name, email and phone arrive already filled in. People submit in seconds, sometimes by reflex, without having read the offer. Volume swells, cost per lead collapses, the dashboard turns green. Three weeks later, your sales reps are calling people who do not remember you.
The only KPI that commits: cost per qualified lead
At DecisionIA, the first sign-ups checked by hand held a surprise: only a fraction matched the intended target. The rest was a mix of curious onlookers, employees with no purchasing power and (more on this below) peers. The arithmetic is merciless: if only one lead in three is on target, a €4 cost per lead is really a €12 cost per qualified lead. The flattering dashboard number has just tripled without a single extra euro being spent.
Hence a rule I apply before any scaling: check a sample of leads by hand. One hour of work, fifty rows, one column marked “on target / off target”. It is the only way to establish the real qualification rate, and therefore the real cost. It is the same principle as calibrating intermediate events, which I detail in my article on Meta Ads for live events: never use a proxy metric without its exchange rate.
Qualify in the creative, since the targeting will not
If Meta cannot filter by job function, the creative has to do it instead. It is a reversal of instinct: in e-commerce you write an ad to maximise clicks; in B2B lead gen you write it to discourage the wrong clicks. Stating the intended audience (“for founders and partners”), the level, and above all the order of magnitude of the price removes noise before the click is even billed. A cost per lead that rises because the creative is filtering is good news.
A second, counter-intuitive lesson: on this client, social proof beat the direct call to action. Creatives built around testimonials and concrete cases generated more clicks and more conversions than “Sign up now”. The reading is simple: on a cold audience that asked for nothing, the direct CTA arrives before any reason to believe. Social proof does the legitimacy work the targeting does not. It is worth testing against audience temperature, but the default hypothesis in cold B2B deserves to be inverted.
Operational corollary: multiply the variants. Four to six distinct angles (testimonial, hard numbers, objection handled, short video) always teach you more about the target than three weeks of audience tweaking.
The form: volume or intent, you have to choose
Meta offers two instant form types, and the choice is structural. The volume-oriented format sends the lead the moment “Submit” is tapped. The Higher Intent format adds a confirmation step: the user reviews their details and has to validate. That deliberate friction removes accidental submissions and anyone not engaged enough to confirm.
Mechanically, volume drops and cost per lead rises. That is the point: in B2B, where a sales meeting costs far more than a click, losing a few dozen percent of leads to eliminate reflex submissions is an excellent trade. The same reasoning applies to the native form versus landing page decision: the native form wins on volume, the landing page wins on intent. The right choice depends on the length of your sales cycle, not on platform habits.
Excluding peers: the niche reflex
This is the parasite specific to specialised sectors. On an AI training offer, targeting the “artificial intelligence” interest attracts AI consultants in droves: curious peers who click and fill in the form, but will never buy. They sell the same thing. A lead that is expensive to acquire and structurally unable to buy is the worst of both worlds.
The fix is simple, but it does not happen on its own: actively exclude those profiles from the targeting, through interests and custom audiences. The raw cost per lead will rise. The cost per qualified lead falls, and that is the only move that counts.
| Setting | Effect on raw CPL | Effect on qualified CPL |
|---|---|---|
| Volume-oriented native form | Sharp drop | Degraded (reflex submissions) |
| Higher Intent form | Rise | Clear improvement |
| Price and audience stated in the creative | Rise | Clear improvement |
| Social proof rather than direct CTA | Drop | Improvement (test it) |
| Excluding peers | Rise | Strong improvement |
| CRM loop + Conversion Leads goal | Rise | Strong improvement |
| Scaling on raw CPL alone | Apparent drop | Fast degradation |
The most expensive mistake
Tripling the budget because the cost per lead is “excellent”. As pressure increases, the algorithm reaches further into the audience, towards even less qualified profiles, since that is where forms get submitted most easily. The raw CPL stays beautiful, the qualification rate collapses, and nobody notices until the sales team reports that the call list is going nowhere. The right reflex: never scale on a number that has not been matched against the CRM. A CPL that doubles from €4 to €8 while the qualification rate goes from 30 to 80 % is a win, not an incident.
Closing the loop: teaching Meta what a good lead is
Everything above stays artisanal as long as the information does not flow back to the platform. Meta offers a setup that B2B advertisers routinely ignore: sending the real status of leads from the CRM through the Conversions API for CRM, and running campaigns on the Conversion Leads goal. When a lead moves to “meeting booked” in the CRM, that information travels back, and the algorithm finally learns to look for those profiles rather than form fillers.
It is not magic, and Meta's documentation is honest about the prerequisites: use instant forms, generate at least 200 leads a month, upload data at least once a day, and target a funnel stage that occurs within 28 days with a conversion rate between 1 and 40 %. Expect three to four weeks before the full benefit. So this is a setup for an established flow, not for a €500 test campaign. Below those thresholds, you close the loop by hand: infinitely better than no loop at all.
The prerequisite, in every case, is clean tagging. A plan of URL parameters set at ad level, consistent from one channel to the next, so the CRM can tell a paid Meta lead from a Google, LinkedIn or organic one. Without it, no euro spent can be attributed, and the whole budget arbitration falls apart. That job belongs as much to my Analytics service as to media buying, and it follows the same logic as server-side tracking: data made reliable before optimisation, never the other way round.
Meta, Google, LinkedIn: each in its place
The gap measured at DecisionIA, €3.73 on Meta against €7.71 on Google, does not say “Meta is better”. It says each channel buys something different: Google buys declared intent, expensive but warm; LinkedIn buys a professional identity, very expensive but surgical; Meta buys attention cheaply from an audience that asked for nothing. In B2B, Meta is therefore the channel of latent demand: the one that reaches the director who is not yet typing your query. It is formidable at that, provided you accept that qualification costs something and pay for it upfront: in creative, in deliberate friction, in exclusions, in measurement. That is the heart of my Meta Ads service and, more broadly, of the budget allocation I run through my SEA service.
In the crucible: distilling volume down to the lead that buys
B2B lead generation on Meta is a work of distillation: the platform delivers an abundant, cheap must, and the whole craft consists of separating the few drops worth something. The lead is the €4 form filled in five seconds by someone who will never buy; the gold is the meeting booked with a decision-maker, even if it cost three times more.
The dashboard will never make that distinction for you: it counts forms, not customers. It falls to the advertiser to bring the measurement, the CRM and the courage to let a reassuring number rise. I work the full chain (creative, forms, exclusions, tagging, CRM loop) on my Meta Ads service page.
Find out what your Meta leads are really worth
I audit your lead gen setup (targeting and exclusions, form type, creative angles, UTM tagging, CRM feedback) and hand you your real cost per qualified lead, channel by channel, with the plan to bring it down.