How a tight-ICP B2B team unlocked Meta—without losing qualification

Company & GTM Profile

"People fixate on the platform. We focus on on-target reach and efficiency. Primer lets us unlock B2C channels for B2B—without losing ICP discipline.” — Colin, Growth Marketing, Zip

Zip runs a narrow ICP with long sales cycles, so every conversion counts. This quarter they doubled to nearly tripled Meta investment after seeing performance gains. Meta’s media costs are often a third of LinkedIn’s, and when targeting stays inside the ICP, down-funnel SQL costs stay comparable  while generating additional scale. That shift lets Zip scale without sacrificing fit.

The Challenge: Diversify channel mix, keep enterprise-grade fit

LinkedIn’s targeting is strong but expensive. Meta offers scale and efficiency—but at the cost of relevance. For a company like Zip, where each deal involves a long, expensive sales cycle, pure lead volume isn’t useful. The goal was to make Meta behave like a B2B channel: broad reach, enterprise precision.

"At the end of the day, it’s on-target reach. Primer lets us get it—on Meta—without giving up our ICP.”

How Zip Makes Meta Work for B2B

Zip runs Meta like a system: Constrain, Signal, and Sustain. Primer is the layer that keeps that system pointed at the ICP.

1. Constrain: Point the Algorithm at the Right Market

Zip runs every Meta campaign from the same starting point — a clean ICP built in Primer  Primer provides enough scale for Meta’s machine learning to optimize while confining its reach to verified accounts inside Zip’s buying zone.

From that shared base, the team runs two distinct regional strategies.

In North America, where competition is dense and fatigue sets in fast, Zip goes inclusion-only. Primer audiences run on their own — no Advantage+, no lookalikes, no expansion layers. This setup preserves quality and ensures Meta learns only from qualified buyers.

In international markets, the dynamic flips. Inventory is broader, so Zip layers Advantage+ on top of Primer audiences. Colin is considering adding Anti-ICP suppression next, to further “box in” the algorithm and eliminate wasted learning cycles.  That blend maintains ICP precision while unlocking additional reach and lower CPMs — a combination that’s delivered the best performance globally so far.

Colin keeps the system in motion. Every few months he re-tests Meta features — like Advantage+ or new optimization modes — to see whether platform updates or market shifts change the outcome. “What didn’t work last quarter might work now,” he says. “Meta evolves fast; you have to keep testing.”

"Meta is incredibly efficient at acquiring leads. The discipline is constraining who it learns from. Primer is how we do that.”

2. Signal: Engineer Feedback the Platform Can Use

Meta optimizes on a conversion-window that's often too short for B2B advertisers. Zip’s conversions can take six to nine months. The challenge is giving Meta short-term signals that correlate with long-term value.

The team runs a dual-signal setup (pixel + CAPI) to ensure Meta receives every conversion reliably. Meta lead forms now validate work emails and they redesigned their forms to ask better qualifying questions, filtering junk before it enters the funnel. Headcount, seniority, and role act as early-stage value proxies, signaling lead quality before true revenue attribution is visible.

"In B2B you can’t feed eight-week signals into a seven-day optimizer. We focus on near-term, high-signal proxies and make them count.”

The result: Meta delivers leads at a fraction of LinkedIn’s cost, while down-funnel efficiency stays even.

3. Sustain: Keep Performance Stable at Scale

With a limited ICP size, freshness is everything. Primer’s continuously refreshed audiences help, but creative rotation matters just as much.

On Meta, Zip runs 12–15 active ads per ad set, compared with 4–6 on LinkedIn, to keep engagement high and avoid fatigue. Colin can do this because Meta’s machine learning optimizes creative delivery faster and at a far lower CPM, allowing multiple ads to compete and self-select the winners without overspending.

Where LinkedIn CPMs make saturation nearly impossible, Meta’s low costs mean frequency builds quickly. Colin watches frequency and qualitative indicators in lieu of penetration reporting—penetration reporting being an area where Meta lags behind.

"If audiences are static, you burn them out. Primer’s refresh and a larger rotating creative set keep Meta learning.”

Why Primer

When comparing vendors, Zip valued transparency, product velocity, and fit for a self-sufficient B2B team.

With Primer, Colin could ask where a specific lead came from and get a clear answer—a level of openness missing from other tools. Primer also continues to evolve, adding measurement and audience testing features instead of standing still.

"We didn’t need a managed-service layer. We needed a partner focused on audience quality—and still building.”

Practical Guidance from Zip

• Better signals beat broader targeting. Feed Meta a clean ICP audience and constrain learning.
• Optimize for quality, not cost. Use form questions and company data as early value proxies.
• Dual-signal setup improves reliability. Fire conversions through both pixel and server-side CAPI.
• Creative freshness sustains reach. Rotate ad formats frequently to avoid fatigue.
• Test, then re-test. Cycle Advantage+ back in as Meta evolves; don’t assume static results. Meta invests in its algorithm.

"It’s not about channels being B2B or B2C,” Colin said. “It’s about on-target reach. Primer helps us get it.”

What’s Next

Zip is preparing to test Anti-ICP suppression within Advantage+ campaigns to control learning even further. They’re also exploring deeper first-party integrations—data warehouse and outreach tools—to create custom audiences directly from enriched internal data.

The Outcome

With Primer, Zip turned Meta from a broad B2C network into a cost-efficient B2B growth engine. The team scaled investment several-fold, cut media costs dramatically, and proved that Meta can drive enterprise-quality leads at LinkedIn-level efficiency.