Buy Attention Where It’s Cheap, Close on LinkedIn: The Feeder-Funnel Arbitrage That Cuts CAC

We stopped paying LinkedIn to prospect. We pay other channels to pre-qualify and only pay LinkedIn to close.

That single shift turned a money sink into an efficient, repeatable CAC engine.

The core problem

• LinkedIn’s targeting is excellent. Its CPMs aren’t.
• Running cold prospecting on LinkedIn pushes CAC up before we’ve proven audience–offer fit.
• Most teams try to “fix” this with bid tweaks or new creatives. The real fix is cost structure.

The idea: Feeder → Closer

• Buy low-cost, contextually relevant traffic on channels with cheaper reach (YouTube, Reddit, Meta, niche newsletters).
• Send that traffic to a short, high-intent content page (not a demo page yet).
• Build segmented Website Audiences in LinkedIn from those visits (by URL rules and UTM source).
• Narrow those audiences further with LinkedIn attributes (job function, seniority, company size).
• Use LinkedIn only for the conversion ask (Lead Gen form, demo, pricing request).

In short: arbitrage attention. Buy discovery where it’s cheap. Pay LinkedIn only when we have proof someone likely fits our ICP.

What you need in place

1) LinkedIn Insight Tag across your site.

2) Clear “high-intent” URLs to segment by. Examples:

• /pricing
• /case-studies/*
• /product/* (features, integrations)
• /solutions/* (role or industry pages)

3) UTM discipline. Append utm_source, utm_campaign, and utm_content to every feeder click. We’ll segment retargeting pools by source and message.

4) A 60–120 second content page that pre-qualifies (short video + skim text). The goal is idea transfer, not lead capture yet.

Step-by-step build

1) Map your ICP slices.

• Example: “Mid-market RevOps leaders in SaaS” vs “Ecommerce CMOs at $50–250M GMV.” Different messages and feeder channels.

2) Pick 2–3 feeder channels per ICP.

• YouTube: contextual placements on videos about your core problem (how-tos, vendor comparisons, teardown content).
• Reddit: targeted subreddits + keyword targeting in threads that match your problem domain.
• Meta: narrow interest stacks + lookalikes from high-LTV customers (keep overlap with LinkedIn acceptable; we’ll filter later).
• Niche newsletters/sponsorships: UTM’d links to your content page.

3) Drive to a single “qualifier” page per ICP.

• Page components:
• Hook: the problem stated in the buyer’s words.
• 60–120 sec video: show outcome, not features.
• 2–3 proof points (logos, one-sentence case study outcomes).
• A soft CTA (optional) — but the primary goal is to tag qualified visits.

4) Build segmented LinkedIn Website Audiences.

• Audience A: URL contains /pricing OR /case-studies → name: “High Intent – Any Source.”
• Audience B: URL contains the qualifier page AND utm_source=youtube → name: “YouTube Qual – ICP1.”
• Audience C: URL contains qualifier page AND utm_source=reddit → name: “Reddit Qual – ICP1.”
• Create separate pools for each ICP and each source. Minimum size matters, so start broad enough to fill.

5) Layer LinkedIn attributes on top of Website Audiences.

• Example: Target “YouTube Qual – ICP1” AND Seniority = Manager+ AND Job Functions = Marketing, Operations, Sales Ops AND Company Size = 200–2,000.
• This is where magic happens: we only pay LinkedIn to show conversion ads to pre-warmed people who also match ICP traits.

6) Build the LinkedIn “closer” campaigns.

• Campaign per ICP per feeder source (keeps budgets and learning clean).
• Objective: Lead Generation or Website Conversions.
• Formats: Single image + Document Ads for proof, Video for product-in-action, and Lead Gen Forms for speed.
• Creative: outcome-first headline, 1–2 social proofs, clear ask (demo, audit, pricing, trial).

Budgets and bidding (start here)

• Split: 60% feeder, 40% LinkedIn closer for the first 3–4 weeks. Adjust after you see conversion-to-meeting rates.
• Feeder CPM/CPC should be 30–70% lower than LinkedIn cold. If not, switch feeder placements.
• LinkedIn bidding: start with Maximum Delivery for retargeting until you hit stable daily conversion volume, then test manual CPC to cap spikes.
• Daily pacing: keep closer campaigns always-on. Let feeder flights pulse if needed to refill pools.

Rule of thumb for rebalancing weekly:

• If LinkedIn closer CTR < 0.6% on warm pools, your qualifier page is attracting the wrong people or the ask is misaligned. Fix message first.
• If pool sizes stall, your feeder reach is too narrow. Add placements/keywords, not broader ICPs.
• If form completion rate < 8–10% on Lead Gen, shorten fields and add a 1-liner “what you’ll get.”

Creative that converts warm traffic

• Swap feature lists for “before/after” copy and timeline (“Cut reconciliation from 3 days to 3 hours in 30 days”).
• Use Document Ads as proof carriers (2–4 page mini-case with a native CTA on the last slide).
• Mirror the qualifier page headline in the LinkedIn ad. Message continuity beats novelty here.
• For SaaS trials: add role-based templates in the ad text (“Start with the RevOps template”). It bumps intent without adding friction.

Measurement that keeps you honest

• Build a blended CAC model with clear swimlanes:
• Feeder spend → “Qual visits” (sessions on qualifier/pricing/case pages)
• LinkedIn closer spend → Leads → SQAs → Meetings → Pipeline → Revenue
• Attribute revenue to the closer channel by default, but track “source of warm” from UTMs to understand which feeder fills the highest-quality pools.
• Create a holdout: 10–15% of warm traffic excluded from LinkedIn closer. Use it to estimate incrementality vs. organic conversions.
• Watch overlap. Deduplicate audiences and set source-priority rules so people don’t see five asks in one week.

Mini case: mid-market SaaS (composite)

• ICP: RevOps leads at 200–1,500 employee SaaS firms.
• Feeder: YouTube contextual + Reddit threads seeded to a “Forecasting Accuracy” explainer page.
• LinkedIn closer: Lead Gen (demo) + Document Ads (2-page case) to Website Audiences segmented by UTM source, narrowed by seniority and job function.
• Outcome after two sales cycles:
• Blended CPL down 35% vs. LinkedIn-cold control.
• Demo-to-SQL rate up 19% (warmer, better-fit traffic).
• CAC down 28% on the RevOps segment; 17% on Sales Ops (message resonance difference showed up fast).

The kicker: we spent less finding people, and more asking qualified people to act.

Common failure modes (and fixes)

• “Our warm pool isn’t converting.”
• Fix: Your qualifier page doesn’t filter. Add a 60–120 sec video that states who this is not for. Tighten the LinkedIn narrowing (function, seniority, company size).

• “Pools are too small to scale.”
• Fix: Expand feeder placements before broadening ICP. Add long-tail YouTube placements, more subreddits, or a newsletter slot with strong topic overlap.

• “Quality dipped when we added Meta as a feeder.”
• Fix: Split by utm_content and only include visits with 2+ pageviews or “/pricing” viewers in your Website Audience. If needed, create a second qualifier page that repels tire-kickers.

• “Attribution is messy.”
• Fix: Maintain a holdout cell. Align on a primary credit rule (closer-first) and a secondary view for feeder influence. Don’t average your way to confusion.

Who should run this (and who shouldn’t)

Run it if:

• You sell to defined B2B roles where LinkedIn has strong attribute filters.
• Your CPMs on LinkedIn cold are choking scale.
• You can produce one strong 60–120 sec explainer per ICP.

Skip it if:

• Your ACV can’t support two-step paid motion.
• You can’t maintain UTM discipline or segmented audiences (this whole play rides on clean pools).

The takeaway

We don’t need LinkedIn to find our buyers. We need LinkedIn to close them.

Buy attention where it’s cheap. Ask for action where precision and trust are highest.

If your LinkedIn cold CAC is creeping up, build one feeder-to-closer track for your top ICP and run it for two sales cycles. If the blended CAC doesn’t beat your control, we’ll audit your setup and share where the leak is.

Want the build deck and audience templates? Drop “Feeder” in a reply and we’ll send the kit.