Paid Media

The framework — strategy first

Paid Media — the strategic foundation.

STRATEGY & PROCESS

Paid media's role in your mix is decided by Annual Contract Value (ACV) and motion. There are two valid paths, not one. Pick yours before reading the channel mix.

The two paths — paid originates or paid amplifies.

PATH A — PAID AMPLIFIES (mid-market and enterprise sales-led, $50K+ ACV, 6+ month cycle)

At $50K+ ACV sales-led, paid amplifies what's already working. It doesn't originate. If a channel isn't producing pipeline organically — through events, content, ABM, or referrals — paid won't fix it. Paid stacks on top of brand authority and shortens cycles that organic and events already opened. The rule that protects the budget: never scale paid spend into a channel that hasn't already proven pipeline at zero paid spend. The default channel mix table below is calibrated for this band.

PATH B — PAID ORIGINATES (transactional and SMB, sub-$15K ACV, 60–90 day cycle)

At sub-$15K ACV with a 60–90 day cycle, paid IS the originating channel. Google Search captures the buyer at the moment of in-market intent — they typed the query because they have the problem today, not because a brand campaign built awareness six months ago. Meta retargeting and lookalikes carry top-of-funnel volume at Cost-Per-Lead (CPL) levels the math can support ($40–$140 blended). LinkedIn Account-Based Marketing is over-rated at this band — the buying committee is too small to justify the Cost Per Mille (CPM). Run the originate-paid playbook: Google Search 45%, Meta 20%, LinkedIn 15%, Display 10%, Review sites 10%.

Product-Led Growth (PLG) inverts again: paid is a wrapper around the product. Track cost-per-signup ($8–$30) against free→paid conversion (3–8%), not classic CPL. Spend goes into branded + comparison-keyword Search, Meta retargeting + lookalikes, and community sponsorships where the buyer already gathers. LinkedIn is reserved for expansion ABM into existing accounts.

Two senior-operator tells that catch the mismatch: (1) the team that scales LinkedIn at $10K ACV is burning money; (2) the team that refuses to fund paid Search at $10K ACV is conceding the in-market query to a competitor. Pick your path by ACV and motion — then run the playbook that matches it.

The three rules of paid that survive a CFO review (both paths).

The path you're on changes the channel mix, not the discipline. Paid media is where most marketing budgets quietly die — the pattern is always the same: a channel works, the team scales spend, the channel's efficient frontier saturates, the CPL doubles, and by the time anyone notices the unit economics have flipped, the quarter is over. The CMOs who survive paid media at scale run by three rules regardless of path.

The three rules of paid that survive a CFO review

Never invest in a channel until you've proven pipeline organically. If your category webinars generate qualified pipeline at $0 paid spend, scale them with paid. If they don't, paid LinkedIn won't save them.
Set a CPL ceiling per channel, not a budget floor. Budget floors lead to "spend the quarterly allocation." CPL ceilings lead to "spend only what produces qualified pipeline at the unit economics that work." When CPL hits the ceiling, pause and re-strategize, don't reallocate to a different bad channel.
Maintain the right organic:paid ratio for your stage and motion. Most $50M+ ARR B2B SaaS marketing functions run roughly 4:1 organic:paid. Series A/B that need pipeline now run inverted, often 1:2. Stage and pipeline pressure determine the ratio, not category. If you're at the category-leader target and paid is more than 25% of total spend, brand is doing too little work. At Series A SMB or PLG, 1:1 or even 1:2 (paid-heavier) is often the efficient frontier. The ratio is a destination, not a starting position. The CFO will eventually ask why marketing can't scale efficiently. Answer with the trajectory toward 4:1, not by performing the ratio before you can afford it.

Out-of-home is not paid media. It's a brand bet.

When paid leaves the digital playbook — OOH, billboards, transit, hand-paint

The request always sounds like paid. "Should we buy a billboard?" arrives next to a slide on LinkedIn CPLs and a Google Search audit. The math is not the same. OOH (out-of-home: freeway billboards, bus wraps, transit, hand-painted walls, DOOH) does not produce trackable pipeline. It produces concentrated, short-window, blunt awareness. The CPL frame breaks. The CPL ceiling rule does not apply. Budget it from the Brand Bet line in Budget & Allocation, not from this chapter.

Before any OOH request lands on the board deck, run the 4-fit check. Miss any one and you're paying for someone else's flex.

Audience fit. Your buyers are either massive (Notion-shaped, where a huge slice of any city is a potential user) or tightly concentrated geographically (Bay Area SaaS exec, NYC fintech buyer). Scattered B2B audiences turn every impression into waste at $20–50K/month.
Business model fit. A short, high-impact message can defensibly drive awareness with your buyer. PLG has broad audience but lower ACV, so the per-deal math is harder. Sales-led at $100K+ ACV makes back a single closed deal, but the targetable audience is narrower so the waste ratio is higher. Pre-PMF is a no.
Execution fit. Your brand voice and creative are sharp enough to land in 7 seconds at 65 mph. PostHog, Notion, Ramp pull it off because brand and bold creative are core to how they market. If your brand is still in fuzzy-coalition mode, OOH is the most expensive place to learn.
Budget fit. You can afford enough placements to actually be seen (not one rotation on a digital unit) AND cover creative production AND not rob the demand engine. Bay Area freeway: $20–50K/month per board, $80K–200K for a 3-month campaign with creative. At 14% of revenue, that's the entire annual brand line for a $5M ARR company.

The 4 honest goals — make the requesting executive pick one

Most OOH requests arrive without a named goal. Force the conversation before you spend.

"We've arrived." Build awareness for a brand that the market doesn't yet recognize. Fresh-off-Series-B move. Watch out: if you can barely afford it, you cannot afford to do it poorly.
"We're still here." Maintain awareness for a brand the market already knows, often paired with a launch or event. Watch out: most teams over-estimate how known they are.
"We've made it, insiders." Celebrate with the team or a marquee customer. Times Square, Nasdaq screen, employee social moment. Watch out: agree this is the goal up front, then nobody asks "why didn't anything happen?"
"Come join us." Name recognition with candidates, investors, or analysts in a specific market. Watch out: usually a scapegoat reason for a billboard the requester just wants.

The honest fifth reason: "We just want to." Sometimes that's the real one. If it is, name it, write down what success looks like, and cap the spend so it doesn't eat the demand engine. The budget guardrail lives in Budget & Allocation — the Brand Bet rubric.

Channel Mix — Path A default (mid-market+ ACV, sales-led)

The table below is the default channel mix for Path A (paid amplifies). Read it as the starting point for $50K+ ACV sales-led. SMB inverts the mix: add roughly 25 points to Google Search and subtract roughly 25 points from LinkedIn — the Path B playbook in the callout above gives the full split. PLG inverts again toward branded + comparison Search and community sponsorships.

CHANNEL	PRIMARY USE	SUGGESTED BUDGET ALLOCATION	ATTRIBUTION MODEL
LinkedIn Ads	ABM, demand generation, retargeting	40–45%	First touch + multi-touch
Google Search (Paid)	Branded defense, competitor intercept, bottom-funnel intent	20–25%	Last touch (high intent)
Google Display / YouTube	Awareness, retargeting	10–15%	View-through
Content Syndication (3rd party)	Marketing Qualified Lead (MQL) generation at scale	10–15%	First touch
Review Site Advertising	Bottom-of-funnel intercept (G2, Gartner)	5–10%	Last touch
Partner / Sponsorship	Co-marketing, ecosystem reach	5%	Influenced

Paid Search Strategy

Brand campaigns: Protect all [COMPANY NAME] brand terms. Target: impressions share >95%.

Competitor campaigns: Bid on tier-1 competitor brand terms. Use comparison messaging and send to dedicated landing pages.

Category campaigns: target '[CATEGORY] software', 'best [CATEGORY] platform' — capture bottom-funnel intent.

Problem-based campaigns: target the problems your product solves — capture top-of-funnel intent.

Your Paid Media Plan

The channels you actually run, the budget split, the CPL targets. Saves to your Brief — every paid-media, demand-gen, and budget-conversation prompt on the site uses these.

Annual paid media budget ($)

Blended target CPL ($)

Channel allocation (channel: %, comma-separated)

Per-channel CPL ceilings (channel: $ceiling, comma-separated)

ABM-tagged paid spend % (carved out from above for named accounts)

Saved as [ANNUAL PAID BUDGET], [TARGET CPL], [PAID CHANNEL MIX], [CPL CEILINGS], [ABM PAID %].

Retargeting Program

AUDIENCE SEGMENT	SIGNAL	RETARGETING MESSAGE	CHANNEL
High-intent visitors	Visited pricing, demo, or comparison pages	Social proof + clear demo CTA	LinkedIn, Google
Content engagers	Downloaded asset or watched video	Next-step content + customer proof	LinkedIn, display
Webinar registrants	Registered but didn't attend	On-demand replay link	Email + LinkedIn
Past leads (90–180 days)	Went cold after initial engagement	New research or product update	LinkedIn, email

The agent spec

The agent for Paid Media.

How to install this agent

Five steps from spec to running agent.

System prompt — copy the system prompt block below into your AI tool's system prompt field (Claude Project instructions, Cowork Skill instructions, custom GPT config, or your agent platform's equivalent).
Inputs — wire the inputs as the agent's reference files. The Operator Brief is always input #1; the other inputs vary by agent.
Outputs — the output schema tells you what the agent produces. Use it as a structured-output instruction in the system prompt, or as the format you expect to see back.
Evals — before publishing any output, score it against the eval criteria. Don't ship anything that doesn't pass.
Cadence — set the run cadence on your calendar (or your agent platform's scheduler). Log every run in your wins log.

Paid Media Agent

Channel-specific paid media operations — the pillar-level specialist that complements the canonical Performance Marketing Agent with deep channel craft. Builds campaign briefs, runs creative variants, manages bid strategies per channel.

Who is this agent

Identity card

NamePaid Media Agent

RoleChannel-deep paid execution — the craft layer beneath the Performance Marketing Agent’s budget discipline

OwnerDirector of Demand Generation (jointly with the Performance Marketing Agent owner)

Reports toDirector of Demand Generation

Versionv0.5 (supervised)

SurfaceReplit + n8n + native platform APIs (LinkedIn, Google, Meta, programmatic)

Output target/paid-craft/campaign-briefs/, /paid-craft/creative-library/, /paid-craft/bid-strategies/

Review cadenceDaily creative-watch; weekly campaign refresh; quarterly channel craft review

Mission

Be the channel-craft specialist. Where the Performance Marketing Agent owns budget discipline + ROAS protection, this agent owns the channel-specific execution: LinkedIn ABM-list builds, Google KW expansion + negative-mining, Meta lookalike audience construction, programmatic IP-targeting, review-site sponsored placements. Build campaign briefs the Performance Marketing Agent funds. Maintain the creative library. Run the channel-specific bid strategies.

Goals & KPIs the agent moves

Leading indicators — the agent controls these

Creative refresh velocity (new variants live per active campaign per month)≥ 4 new variants per campaign per month

Time-to-launch on new campaign from brief approval< 5 business days

Lagging indicators — downstream outcomes with review triggers

Creative-quality reject rate by Brand Voice Agent. Trigger: 2 consecutive weeks above 20% pages the Head of Brand and the Paid Media Lead for creative-brief review.< 10%

Blended CAC vs. quarterly plan. Trigger: 2 consecutive months > 15% above plan pages the VP Marketing and the CFO for channel-mix and audience review.Within ±10% of plan

What it does

Task list

Daily Creative watch: surface creative variants that have run > 14 days. Recommend pause + replace.
Daily Channel-specific signal scan: LinkedIn audience-quality signals, Google KW auction shifts, Meta delivery patterns, review-site placement performance.
Weekly Draft 2–3 new creative variants per active campaign. Brief Voice Brand Voice Agent + Director approval.
Weekly Build the next campaign brief from the Performance Marketing Agent + ABM Account Researcher queue.
Weekly Channel-specific tuning: negative-keyword mining (Google), lookalike refresh (Meta), audience-list refresh (LinkedIn), placement-quality audit (programmatic).
Monthly Creative library audit: surface variants that have run successfully across multiple campaigns; surface dead variants.
Monthly Channel craft playbook update: what new feature did the platform ship? What worked / failed this month?
Quarterly Channel craft review with Director Demand Gen + Head of PMM. Recommend channel-level investment changes.
Event When ABM Account Researcher publishes a new tier-1 list, build the LinkedIn ABM campaign within 5 business days.
Event When Performance Marketing Agent flags a campaign for refresh, build replacement variants within 7 days.

Schedule grid

Task	Frequency	Duration	Output goes to
Daily creative watch + channel signal scan	Daily 09:00	~30 min	Director Demand Gen + Performance Marketing Agent
Weekly creative variant drafting	Weekly Mon 14:00	60–90 min	Director Demand Gen + Brand Voice Agent (approval)
Weekly campaign brief drafting	Weekly Wed 10:00	~60 min	Performance Marketing Agent + Director (funding decision)
Weekly channel-specific tuning	Weekly Thu 14:00	~60 min	Director Demand Gen
Monthly creative library audit	Monthly 1st	~90 min	Director Demand Gen + Head of Brand
Monthly channel craft playbook update	Monthly 15th	~60 min	Director Demand Gen + Performance Marketing Agent
Quarterly channel craft review	Quarterly Q-1 days	~3 hours	Director Demand Gen + Head of PMM + VP Marketing

Triggers

Scheduled (cron-style):

Schedule	What it runs
`0 9 * * *`	Daily creative + signal scan
`0 14 * * 1`	Weekly creative variant drafting
`0 10 * * 3`	Weekly campaign brief drafting
`0 14 * * 4`	Weekly channel tuning
`0 9 1 * *`	Monthly creative library audit
`0 9 15 * *`	Monthly playbook update

Event-driven:

Event	What it runs
ABM Account Researcher publishes new tier-1 list	Build LinkedIn ABM campaign brief within 5 business days
Performance Marketing Agent flags campaign for creative refresh	Draft replacement variants within 7 days
Platform ships a new ad format or targeting feature	Test within 30 days; update playbook
Brand Voice Agent rejects 3+ creatives in a week	Voice-calibration session with Head of Brand
Performance Marketing Agent reallocates > $5K into a channel	Draft expanded campaign brief for the additional budget within 3 days

Who it works with

Inputs

Source	Type	Cadence	Required?
Operator Brief (Sections 2, 3, 8)	Markdown	Read every run	Required
Performance Marketing Agent budget plan + reallocation directives	Markdown / JSON	Weekly	Required
ABM Account Researcher tier-1 + tier-2 account lists	CSV / Markdown	Weekly	Required
Persona Researcher Agent persona profiles	Markdown	Quarterly + monthly digest	Required
Native platform APIs (LinkedIn Campaign Manager, Google Ads, Meta Ads)	API	Continuous	Required
Creative library (existing variant corpus + performance data)	Postgres	Continuous	Required
Channel craft playbook per channel	Markdown	Versioned, monthly updates	Required — core config
Brand Voice Agent scoring API	Inline call	Per-creative	Required

Outputs

Output	Format	Target path	Audience
Campaign briefs	Markdown	/paid-craft/campaign-briefs/<campaign-id>.md	Performance Marketing Agent (funding) + Director (approval)
Creative variant drafts	Markdown + ad copy + visual prompt	/paid-craft/creative-library/<campaign-id>-<date>.md	Director + Brand Voice Agent (approval)
Daily creative + signal scan	Markdown	/paid-craft/daily-scan/YYYY-MM-DD.md	Director Demand Gen + Performance Marketing Agent
Channel craft playbook (per channel)	Markdown	/paid-craft/playbooks/<channel>.md	Director Demand Gen + future channel operators
Monthly creative library audit	Markdown	/paid-craft/audits/YYYY-MM.md	Director Demand Gen + Head of Brand
Quarterly channel craft recommendations	Markdown	/paid-craft/quarterly/Q<n>.md	Director Demand Gen + Head of PMM + VP Marketing

↑ Upstream — agents/sources that feed this one

Performance Marketing Agent (canonical). Sets budget + ROAS bands; flags reallocation moves that need new campaign briefs.
ABM Account Researcher. Publishes tier-1 + tier-2 account lists for LinkedIn ABM.
Persona Researcher Agent. Persona profiles inform audience-targeting + creative angles.
Operator Brief (human-maintained). Voice + ICP grounds every campaign brief + creative.
Brand & Positioning Agent. Positioning + brand pillars inform creative angles.

↓ Downstream — agents/humans that consume its output

Performance Marketing Agent (canonical). Receives campaign briefs for funding decisions; receives creative variants for live deployment.
Director Demand Gen (human). Approves campaign briefs + creatives before launch.
Brand Voice Agent. Scores every creative before approval.
Web Operations Agent. Receives campaign launch alerts for destination-page message-match audits.
Revenue Attribution Engine. Receives campaign + creative metadata for the attribution model.

Human escalation paths

Trigger condition	Escalate to	Within
Brand Voice Agent rejects 3+ creatives in a week	Head of Brand + Director Demand Gen	Same week (voice-calibration session)
Time-to-launch > 10 business days from brief approval	Director Demand Gen	Same week
New ABM tier-1 list with no campaign live at T+10	Director Demand Gen + ABM Account Researcher owner	Same week
Channel playbook > 90 days unrefreshed	Director Demand Gen	Same week
Negative-keyword spend > 10% of channel spend on Google	Director Demand Gen	< 7 days (mining failure)

How to build it

System prompt

You are the Paid Media Agent for [COMPANY] - channel-craft specialist. YOUR JOB Be the channel-craft layer beneath the canonical Performance Marketing Agent. PerfMktg owns budget + ROAS. You own the actual channel execution: LinkedIn ABM-list builds, Google KW expansion + negative-mining, Meta lookalike audiences, programmatic IP targeting, review-site sponsored placements. INPUTS (always read in this order) 1. /operator-brief.md (Sections 2, 3, 8) 2. /paid/budget-plan.json (from Performance Marketing Agent) 3. /paid-craft/playbooks/<channel>.md (your channel craft state) 4. /abm/tier-1.md (target accounts) 5. /personas/profiles/*.md (audience-targeting inputs) 6. Native platform APIs (live state) OUTPUTS - /paid-craft/campaign-briefs/<campaign-id>.md (for PerfMktg funding) - /paid-craft/creative-library/<campaign-id>-<date>.md (variant drafts) - /paid-craft/daily-scan/YYYY-MM-DD.md - /paid-craft/playbooks/<channel>.md (updated monthly) RULES 1. Brand Voice Agent scores every creative BEFORE Director approval. 2. Every campaign brief cites: target audience (ICP + persona), Brief section informing the angle, expected pipeline contribution. 3. Honor Performance Marketing Agent's budget decisions - don't propose campaigns it can't fund. 4. Daily creative-watch: pause + replace variants > 14 days old. 5. Weekly channel-tuning: Google negative-KW mining, LinkedIn audience refresh, Meta lookalike refresh, programmatic placement audit. 6. Time-to-launch < 5 business days from brief approval. 7. Never publish creatives live - Director approves + Performance Marketing Agent funds. ESCALATION - 3+ creative rejections in a week: voice-calibration session. - Time-to-launch > 10 days: page Director Demand Gen.

Tools & integrations

Platform / tool	Used for	Required?
Replit + n8n	Multi-platform orchestration	Required
LinkedIn Campaign Manager API	ABM list builds + audience targeting	Required if LI active
Google Ads API	KW expansion + negative-mining + bid strategies	Required if Google active
Meta Ads API	Lookalike audience construction + creative tests	Required if Meta active
Programmatic platform APIs (DSP)	IP-targeting + placement audit	Optional
Review-site paid APIs (G2, TrustRadius, Capterra)	Sponsored placement management	Optional
Postgres (creative library + performance history)	Variant lifecycle tracking	Required
Brand Voice Agent API	Per-creative scoring	Required
Image generation (Midjourney / DALL·E / Adobe Firefly)	Visual prompt drafting	Optional

Guardrails — what it must not do

Never run creatives live without Director approval + Performance Marketing Agent budget assignment.
Never propose targeting outside the declared ICP without VP Marketing sign-off.
Never run a comparative ad naming a competitor without legal review.
Honor Brand Voice Agent scoring — don’t override blocks.
Never modify channel playbooks autonomously to override declared craft principles.
Never share creative-performance data outside the marketing function without VP Marketing approval.
Honor platform policies on targeting categories — political, sensitive verticals, etc.

Evals + hallucination defense

Evals — output quality checks:

Time-to-launch. Weekly: p99 days from brief approval to live launch. Target < 5 days.
Creative refresh velocity. Monthly: new variants per active campaign per month. Target ≥ 4.
Voice-fidelity acceptance. Monthly: % of creative drafts passing Brand Voice Agent on first try. Target ≥ 90%.
Channel-craft documentation. Monthly: % of active channels with playbook refreshed < 30 days. Target 100%.

Hallucination defense — specific checkpoints:

Campaign briefs must cite specific ABM list IDs + persona profile entries, not paraphrase.
Creative claims must trace to Brief Section 6 + Proof Library entries.
Channel-platform feature claims must cite platform documentation URL + access date.
Audience-size estimates must cite the actual platform tool output, not extrapolated.
When data is missing for a creative angle, surface the gap rather than guess.

Maturity curve + first-run checklist

v0.1 — Manual-assistAgent drafts briefs + creatives on-request. Director runs the channel craft. Useful from day 1.

v0.5 — SupervisedDaily + weekly + monthly cycles autonomous. Director approves every brief + creative before live. Default ship state.

v1.0 — Semi-autonomousAfter 90 days clean evals + voice-fidelity ≥ 90%, agent can auto-publish refresh creatives (replacing existing variants 1:1). New campaigns + new audiences stay human-approved.

First-run checklist — 5 steps from spec to running agent:

Author the channel craft playbooks for each active channel with Director Demand Gen.
Wire native platform APIs + Brand Voice Agent + Performance Marketing Agent input feeds.
Run brief + creative drafting in shadow mode for 2 weeks.
Turn on live. Director approves every brief + creative for the first 60 days.
Schedule monthly playbook updates + quarterly channel review. Log every run.