Four layers decide whether ad spend becomes pipeline: paid media captures demand, creatives convert it, the landing qualifies it, and the CRM records what closed. Most teams run the four as separate projects, so the lead falls between them. Wire them as one flow and every dollar traces back to a contract.
The ads are running. The report looks healthy: impressions up, CPC down, a few hundred form fills. Then sales asks which campaign produced last quarter's three closed deals, and nobody in the room can answer.
That question is the whole test. A team that can't answer it ends up optimising for whatever the dashboard happens to show, which in paid media means clicks. Clicks are the cheapest thing to buy and the loosest proxy for revenue there is.
A B2B paid media system, or acquisition system, is four connected layers: paid media captures demand, creatives convert it, the landing qualifies it, and the CRM records which of it turned into revenue. Each layer feeds the next, and the CRM feeds the ad platforms back. Paid media is only one of the four, and run on its own it's the layer that produces the least.
Four layers, one flow
What a B2B paid media system is: the four layers that turn ad spend into traceable pipeline.
Campaigns produce clicks. A system produces customers on a schedule you can plan against. The difference between the two is whether the four layers talk to each other.
- Paid media is the reach. The channels that capture demand that already exists, and create demand that doesn't.
- Creatives are the message. What the buyer sees first, built to convert rather than to win awards.
- The landing is the filter. One page per segment, built to qualify rather than to describe.
- The CRM is the memory. It measures, connects and optimises the whole funnel, and it's the only layer that knows what actually closed.
Paid media attracts, creatives convert, the landing qualifies, the CRM closes the loop and tells the ad platforms what a good lead looked like. Break any one of the four and the lead gets lost between silos.
Four layers, one flow. Run them as four separate projects and you're buying clicks.
Your ICP is a problem, not an industry
How to define a B2B ICP that paid media targeting can actually act on.
"Fintech companies in Mexico" is a filter, not an ICP. Two companies on that list can have opposite reasons to buy, or no reason at all. Targeting by industry is how budget goes to accounts that were never going to convert.
Your own data already shows the pattern. The useful question isn't who could use the product. It's who already gets real value from it and carries the highest operational friction today. Four signals to look for:
- Critical operational friction. The failure your product prevents carries a direct cost: audit exposure, downtime, manual rework.
- Data maturity. They already treat missing visibility as a business problem, not a technical one.
- A validated decision cycle. The buyer has authority to implement without routing it through an eight-person committee.
- Compliance urgency. Regulation or security turns your product from nice to have into a requirement.
You're not trying to reach the whole market. You're looking for the segment where CAC is lowest because the pain is sharpest.
One channel per demand state
How to choose between Google Ads, Meta and LinkedIn for a B2B funnel, and in what order to turn them on.
Channels aren't interchangeable. Each one reaches a buyer in a different state of awareness, and picking by default (Google and Meta, because everyone runs Google and Meta) is how budget lands on the wrong part of the funnel.
At any given moment, only a small share of your market is actively shopping. The 95-5 rule from the LinkedIn B2B Institute puts it at roughly 5% in-market at a time. Reaching the 5% and reaching the other 95% are two different jobs, and they take two different channels.
Search intent is where the cheapest qualified lead lives, so that's the floor. It's also finite: there are only so many people typing your category into Google this month. Meta is how you scale past that ceiling, by interrupting buyers who have the problem and aren't looking yet. Awareness comes last, once the layers below it prove they convert.
Channel count matters less than precision. For an ICP whose decision maker isn't scrolling Meta, LinkedIn targeting by title, industry and company size does a job the other two can't, through Lead Gen Forms, Document Ads and Conversation Ads. The mix gets defined by the business, not by habit.
| Channel | Buyer state | CAC | Volume | Qualification | Turn it on |
|---|---|---|---|---|---|
| Google Search | Searching for a solution now | Lowest | Capped by search volume | Highest, intent is self-declared | First. It's the efficient floor |
| Retargeting | Already engaged, hasn't converted | Low | Bounded by traffic you already have | Inherits the source that produced the visit | As soon as there's traffic worth recovering |
| Meta | Has the problem, isn't searching | Mid | High | Looser, the creative does the filtering | Once Search converts and hits its ceiling |
| Has the problem, reachable by job title | Highest per lead | Low | Highest on firmographics | When the decision maker isn't on Meta | |
| Awareness | Doesn't know the problem exists | Not directly attributable | Highest | None at the click | Last, once the layers below it convert |
Half the result is the creative
What makes a B2B ad creative convert, and how many variants are worth testing.
Nielsen and NCSolutions put creative at 49% of incremental sales, the single largest driver of ad performance by a wide margin. In B2B it weighs more than that: the decision cycle is long, the buyer is skeptical by default, and most of the category still runs ads that look like corporate brochures.
- The hook decides everything. The first two or three seconds are the filter. If the scroll doesn't stop there, the rest of the ad doesn't exist. So don't test one hook. Test several and let the data pick.
- Video over static for prospecting. It carries more information in less time, and it's the format the feed rewards.
- One ad, one problem, one ICP. A message written for every prospect converts none of them.
- Proof in every piece. Client results, real numbers, named cases. Without evidence there's no trust, and in B2B there's no conversion without trust.
Testing several hooks per batch is a production problem before it's a media problem. If the creative team can't ship variants fast enough, the media plan quietly shrinks to whatever fits: one ad, one message, no test. That's the bottleneck the producer-grader workflow exists to remove.
Fewer creatives, better aimed. Volume of assets was never the strategy.
The landing does the qualifying
What a B2B landing page and lead form should do before sales ever sees the lead.
The typical flow drops a lead into a generic form on a page written for everybody. The click was specific. The page isn't. That gap is where intent dies, and it's the cheapest thing on this list to fix.
- One page per segment. A single CTA and a message that matches the ad that produced the click.
- A form that qualifies. Every field maps to a CRM property. Sales receives data ready to close, not questions to ask.
- Proof above the fold. Volume, logos, cases. The credibility you already have, working in the first scroll instead of on page three.
- An immediate next step. Delete "we'll be in touch". On submit: a calendar or a WhatsApp thread, right there.
Turning visits into records was never the point. The page is the commercial team's first filter: it either shortens the cycle or hands sales a list to triage by hand.
Attribution or guessing
How to attribute a closed B2B deal back to the ad campaign that produced it.
This is the layer teams skip, and it's the one that makes the other three measurable. Without it the ads run and nobody knows what worked, so every optimisation decision is a guess with a budget attached.
- UTM persistence. A custom script captures UTMs, gclid and fbclid and persists them for 30 days. Every lead carries the exact source that produced it, even when the visit and the form fill happen weeks apart.
- Automatic lead scoring. A lead matching the priority profile fires an alert to sales the moment it lands. Zero manual triage.
- Segmentation and routing. Each audience enters its own flow (sales, nurture, onboarding) from first contact.
- Offline conversions back to the platforms. Closed Won returns to Meta and the Google Ads API. From there the algorithm optimises toward customers instead of clicks.
That last one is the difference between a platform that finds you cheap form fills and one that finds you buyers. Ad algorithms optimise for whatever signal you send them. Send them Closed Won.
If Closed Won never travels back to the ad platform, you're paying the algorithm to find you clicks.
Validate before you scale
How much time and budget it takes to validate a B2B paid media channel before scaling it.
A big budget is a bad way to learn. Four to eight weeks with real benchmarks tells you what a bigger number can't, and it costs a fraction of the education.
- A defined window. Four to eight weeks, with benchmarks agreed before the spend starts: CPM and CTR for ad efficiency, CPL and CVR for cost and conversion, and the signal that actually matters, purchase intent.
- Three scenarios up front. Optimistic, base, conservative. The client sees the no-go before the money moves.
- Decoupled risk. If one channel stalls, the plan doesn't. There's always a second track live from day one.
Then the cadence is simple: cut the losers, scale the winners, review every two weeks with the budget actively managed instead of set and forgotten.
The build itself runs in three phases:
- Build the base. Audit of assets and funnel, CRM setup, tracking infrastructure from UTM through to the conversions API, the first creative batch, landings live by segment, campaigns launched.
- Validate. Budget per channel, CPL, qualification rate, CVR, message testing by ICP, and MQL to SQL by channel.
- Scale what won. Pause the rest, build lookalikes on qualified leads rather than raw form fills, and turn the whole thing into a playbook you can replicate in a new market.
You don't need us to run this. The shape is the valuable part, so take it:
- Wire the four layers before adding budget. Reach, message, filter, memory.
- Pick channels by demand state. Start at the bottom of the funnel and work up.
- Aim fewer creatives at sharper problems. One ad, one pain, one buyer.
- Make the landing qualify. Every field maps to a CRM property.
- Send Closed Won back to the platforms. Otherwise you're optimising for clicks.
If the ads are running and nobody can name which campaign produced last quarter's closed deals, the budget isn't the problem. The wiring is. The rest of how we build this sits in our resources.
FAQ
Why do my B2B ads generate leads but no pipeline?
Usually because the layers underneath the ads aren't connected. The click lands on a page written for everyone, the form doesn't qualify, and the CRM never records which campaign produced the deal that closed. The campaign looks healthy inside the ad platform and produces nothing sales can use.
How much budget do I need to validate a B2B paid media channel?
Less than most teams assume. A four to eight week window with benchmarks agreed before the spend starts (CPM and CTR for ad efficiency, CPL and CVR for cost and conversion, plus real purchase intent) tells you whether a channel works. Scale comes after the signal, not before it.
Should a B2B company start with Google or Meta?
Start with Google Search if there's active demand on commercial-intent keywords. It's the most efficient floor: lowest CAC, highest qualification. It's also finite. Meta is how you scale past it, by reaching buyers who have the problem but aren't searching for a solution yet.
When should a B2B company use LinkedIn Ads instead of Meta?
When the decision maker can be named by job title but isn't going to be reached scrolling a consumer feed. LinkedIn targets by title, industry and company size, through Lead Gen Forms, Document Ads and Conversation Ads. It carries the highest cost per lead of the three and the highest firmographic qualification, so it earns its budget on ICPs where a wrong lead is expensive to process.
How do I send Closed Won deals back to Meta and Google Ads?
Through offline conversion imports: the CRM sends the closed deal, with the click identifier it was captured with, back to Meta and to the Google Ads API. The ad algorithm then optimises toward the profile of buyers who actually signed rather than toward whoever fills in forms most cheaply. If Closed Won never travels back, you're paying the platform to find you clicks.
How do I keep UTMs when the visit and the form fill happen weeks apart?
Persist them client-side instead of reading them at submit time. A script captures the UTM parameters plus gclid and fbclid on the first visit and stores them for 30 days, so the lead carries its real source even when the form fill comes from a later, direct visit. Without persistence, most B2B leads get attributed to "direct" and the campaign that produced them looks like it produced nothing.
What should a B2B landing page form ask?
Only what qualifies the lead, and every field should map to a property the CRM already has. The form is the first filter of the commercial team, not a data collection exercise: it should let sales arrive at the call knowing the segment, the use case and the urgency. Pair it with one CTA per segment, proof above the fold, and a calendar on submit instead of "we'll be in touch".
How do I define a B2B ICP for paid media?
By problem, not by industry. "Fintech companies in Mexico" is a targeting filter, and two companies on that list can have opposite reasons to buy. Look at your own closed deals for four signals: critical operational friction with a direct cost, data maturity, a decision cycle short enough to sign, and compliance or security urgency. The segment where the pain is sharpest is the segment where CAC is lowest.
Start with the audit
Every proposal we write starts with a diagnosis of the funnel that already exists. On the acquisition side: channels exploited versus real lead potential, the current creative and its dominant format, and whether the message speaks to one audience or to all of them at once. On the conversion side: the landing and its estimated leak, the CRM integration (do leads arrive qualified and routed?), and whether the proof you already have is visible or buried.
The diagnosis defines the real bottleneck. Proposing without auditing is proposing the obvious.
Send us access to your ad account and your funnel and we'll return the diagnosis: where the lead leaks, what it's costing, and the first three fixes ranked by impact. Free.
Request the audit →For teams that want the system built rather than diagnosed, that's the engagement: four to eight weeks to validate on a controlled budget, then scale on data instead of on a deck.
Book a call →