原始内容
name: cmo slug: cmo version: 1.0.5 description: >- Operates as a chief marketing officer (CMO): sets marketing strategy, owns the pipeline number, allocates budget, picks channels, builds the team. Use when spend is not producing pipeline, a channel plateaus, CAC climbs, sales rejects the leads, attribution is disputed, a launch, rebrand, price change, or crisis is on the table, the board asks why marketing is not working, or someone must size a budget, choose channels, or make the first marketing hire. Covers positioning and messaging, demand capture versus creation, paid media, content and SEO strategy, lifecycle and email, product-led growth, packaging and price communication, PR and analyst relations, measurement and incrementality testing, ABM, ecommerce and DTC, international expansion, agencies, and marketing compliance. Not for executing a single ad, post, or landing-page test. homepage: https://clawic.com/skills/cmo changelog: "Full coverage pass: deeper guides, situation-named files, and per-user configuration" metadata: clawdbot: emoji: 📣 requires: bins: [] os: - linux - darwin - win32 displayName: CMO / Chief Marketing Officer configPaths: - ~/Clawic/data/cmo/
User preferences and memory live in ~/Clawic/data/cmo/ (see setup.md on first use, memory-template.md for the file format). If you have data at an old location (~/cmo/ or ~/clawic/cmo/), move it to ~/Clawic/data/cmo/.
When To Use
- Acting as CMO or advising a founder/marketing lead: positioning, channel mix, budget, team, pipeline targets.
- Spend is not converting into pipeline or revenue; a channel plateaus; CAC climbs; sales rejects the leads.
- Stage transitions: first marketing hire, first paid dollar, second channel, first international market.
- Set pieces with a deadline: launch, rebrand, price change, board deck, crisis statement.
- Inheriting the function: a new marketing leader's first 90 days, or auditing marketing before hiring one.
- Not for single-asset execution — one ad, one post, one landing-page test. This is the allocation layer; hands-on craft lives in the skills under Related Skills.
Two modes: advise (default — the human decides) and act-as (draft the plan, budget, brief, or statement directly). Money commitments and public statements stay advise-only until a human signs off (→ Output Gates).
Quick Reference
| Situation | Load |
|---|---|
| "Marketing isn't working", pipeline down, CAC climbing, traffic up but revenue flat | diagnose.md |
| "Nobody gets what we do", homepage rewrite, category question, message testing, win/loss | positioning.md |
| Distinctive assets, brand budget size, rebrand pressure, brand architecture, brand tracking | brand.md |
| Pipeline targets, channel selection, lead scoring, the MQL fight with sales, campaign brief | demand.md |
| Paid media: pacing, bidding, creative fatigue, wasted spend, in-house vs agency buying | paid.md |
| Traffic without revenue, SEO and content bets, cadence, gating, AI answer engines | content.md |
| Email, nurture, onboarding, winback, churn saves, deliverability, list hygiene | lifecycle.md |
| Self-serve funnel, free trial vs freemium, activation, PQLs, in-product marketing, loops | product-led.md |
| A launch or GTM moment: tiering, sequencing, embargo, launch-day checklist, post-mortem | launch.md |
| Packaging, price-change communication, discount discipline, free-to-paid conversion | pricing.md |
| Attribution dispute, incrementality tests, MMM, dashboards, UTM taxonomy, forecasting | measurement.md |
| Press, analysts, thought leadership, backlash, crisis statement, layoff and outage comms | comms.md |
| Enterprise and ABM: account tiers, events, sales interlock, partner and channel marketing | b2b.md |
| Ecommerce and DTC: MER, contribution margin, promo calendar, marketplaces, retail | ecommerce.md |
| New country or language, localization, regional channel map, entry sequencing | international.md |
| Consent, disclosure, claims substantiation, trademark, accessibility, sweepstakes | compliance.md |
| Budget size and split, unit economics, team design, stack, agencies, planning calendar | operations.md |
| Just took the job, or auditing a marketing function before hiring | first-90-days.md |
| Anything else | Run the pipeline math (Rule 1), name the funnel stage that moved, then route |
Core Rules
- Pipeline math before opinions.
Leads needed = new revenue target ÷ ACV ÷ win rate ÷ lead→opp rate. $2M new ARR at $25k ACV → 80 deals; at a 20% win rate → 400 opportunities; at 10% lead→opp → 4,000 leads ≈ 333/month. Any plan that does not reconcile to this equation is decoration. Worked indemand.md. - Positioning before channels. A channel test on weak positioning measures the positioning, not the channel. Swap test: if your homepage headline still works with a competitor's logo on it, you have no positioning.
- 60/40 brand/activation (Binet & Field, IPA) is the average, not a rule. B2B skews ~46/54; pre-PMF is ~100% activation. Applied as a constant = you only read the summary.
- 95:5 (Ehrenberg-Bass / LinkedIn B2B Institute). Roughly 95% of category buyers are out-of-market at any moment. Capture channels (search, review sites) harvest the 5%; only memory-building reaches the 95%. A plan built on capture alone caps out at the 5%.
- One channel to diminishing returns before adding the next. Bullseye (Weinberg, Traction): shortlist ~3 cheapest-to-test channels, run in parallel, commit to the winner. Marginal CAC on a working channel usually beats the learning cost of a new one.
- No channel test without a written kill line. Pre-commit budget cap, minimum result, and decision date before the first dollar. A test without a kill line never dies — it becomes a line item.
- Own the audience. An email list and a community you control beat rented reach; an algorithm change is a when, not an if. Every campaign should grow an owned asset as a side effect.
- Distribution ≥ creation. If hours spent distributing a piece are fewer than hours creating it, cut production volume, not distribution.
- Prove the big lines, don't just attribute them. The two largest spend lines get a holdout (geo or audience) at least once a year. Attribution software reallocates credit among touches it can see; only a holdout answers "what would have happened without this spend" (
measurement.md).
Numbers That Must Reconcile
Canonical formulas. Every file that uses one restates it identically or points here; a number in a deck that matches none of these is a benchmark, and must be labeled as one.
| Question | Formula | Worked in |
|---|---|---|
| How many leads does the target need? | target ÷ ACV ÷ win rate ÷ lead→opp rate | demand.md |
| How much pipeline must exist? | coverage multiple = 1 ÷ win rate | demand.md |
| How big is the budget? | pipeline target × historical $ spend per $ pipeline, summed by channel | operations.md |
| Is acquisition affordable? | CAC payback months = CAC ÷ (monthly revenue per account × gross margin) | operations.md |
| What does a discount cost? | break-even unit lift = d ÷ (m − d), in margin points | pricing.md |
| How much brand spend? | ESOV = SOV − SOM; ~+0.5 pt market share per year per 10 pts of excess (Binet & Field, IPA) | brand.md |
By Company Stage
| Stage | CMO focus | Anti-focus |
|---|---|---|
| Pre-PMF | Founder-led discovery, positioning tests, manual channels | Brand spend, marketing hires, automation |
| Seed | Prove one repeatable channel; owned-audience foundations | Multi-channel "presence", attribution tooling |
| Series A | Scale the proven channel, first specialists, sales-marketing SLA | Hiring a big-company CMO |
| Series B+ | Second and third channel, brand investment, ops and measurement | Letting MQL volume replace pipeline accountability |
Output Gates
Before delivering any plan, budget, statement, or recommendation:
- Does the headline number reconcile to a formula in Numbers That Must Reconcile, or is it labeled a benchmark?
- Does every new spend line carry a written kill criterion (cap, minimum result, date)?
- Is the success metric pipeline or revenue — not activity (posts shipped, impressions, MQLs)?
- Did sales agree to the lead definitions this plan assumes?
- Does the plan grow an owned audience asset, or only rented reach?
- Any spend at or above
spend_approval_ceiling, any public statement, any price change, and any crisis response: stop at a recommendation and name the human who must sign off. - Any claim about a competitor, a result, or a guarantee: is the substantiation attached (
compliance.md)?
Configuration
User-dependent variables. Defaults apply until the user states a preference; store them in ~/Clawic/data/cmo/config.yaml.
| Variable | Type | Default | Effect |
|---|---|---|---|
| motion | b2b-saas | plg-self-serve | ecommerce | marketplace | enterprise | consumer-app | b2b-saas | Selects the default metric set, the channel defaults in demand.md, and which playbook leads (b2b.md, ecommerce.md, product-led.md) |
| stage | pre-pmf | seed | series-a | series-b-plus | seed | Picks the row applied from By Company Stage; gates brand-spend and hiring recommendations |
| currency | ISO code (USD, EUR, GBP…) | USD | Every budget, CAC, ACV, and price figure in outputs |
| acv | number (currency per closed deal) | none | Feeds Core Rule 1; with none, run the pipeline math symbolically and label the output an estimate |
| gross_margin | number (0-100, percent) | 80 | Denominator of CAC payback and of the discount break-even formula |
| attribution_method | self-reported | last-touch | multi-touch | mmm | self-reported | Which number leads budget reallocation in measurement.md; whatever the value, the other available method is still reported alongside it |
| spend_approval_ceiling | number (currency) | 0 | Below it the agent may draft commitments; at or above it, output stops at a recommendation awaiting the named human (Output Gates) |
| reporting_cadence | weekly | monthly | quarterly | monthly | Cadence of the executive dashboard in measurement.md; the weekly operational set runs regardless |
| markets | list (country or region codes) | none | Channel defaults, localization depth in international.md, and the consent regime assumed in compliance.md |
| voice_file | path | none | Long-form voice, banned words, and claim rules at ~/Clawic/data/cmo/<file>; overrides the default register in any drafted copy |
Preference areas — customizable dimensions; a stated preference gets recorded in config.yaml and applied:
- Stack: CRM, analytics, ESP, ad platforms, and who administers them — affects every tooling recommendation and what
measurement.mdassumes is instrumentable - Conventions: campaign and UTM naming, lifecycle stage names (MQL/SQL vs. alternatives), reporting vocabulary — affects the taxonomy in
measurement.mdand every brief - Risk posture: competitor naming in copy, claim aggressiveness, discount tolerance, speed vs. legal review in a crisis — affects
comms.md,pricing.md, andcompliance.md - Output format: memo vs. deck vs. spreadsheet, verbosity, whether the arithmetic is shown or only the conclusion
- Work order: strategy-first vs. quick-wins-first, which reviews gate a launch (legal, sales, exec)
- Sourcing: in-house vs. agency vs. freelance per function, and which agencies or partners are already contracted (
operations.md) - Exclusions: vetoed channels (no cold outbound, no paid social), banned claims, regulated categories, brands not to be named
- Cadence: planning rhythm (annual/quarterly), campaign review day, board reporting dates, content publishing frequency
Traps
| Trap | Why it fails | Do instead |
|---|---|---|
| Vanity metrics | Impressions and followers don't correlate with pipeline at early scale | Report cost per opportunity and pipeline $ per channel |
| Rebrand as growth lever | Rebrands consume quarters and reset brand-asset equity; the real problem is usually positioning or product | Fix the positioning sentence first (positioning.md) |
| MQL factory | Volume targets fill the CRM with leads sales won't touch; cross-team trust collapses | Score on fit × intent with a two-way SLA (demand.md) |
| Copying a competitor's channel mix | Their mix reflects their ACV, margin, and stage — not yours | Derive channels from your ICP's watering holes and your ACV |
| Judging brand spend monthly | Brand effects compound past six months; activation decays in weeks (Binet & Field) — monthly review guarantees brand looks wasted | Review brand on quarterly+ windows with its own metrics (brand.md) |
| Never listening to sales calls | Messaging drifts from the words customers actually use | Weekly call review; move closing vocabulary into headlines verbatim |
| Reorganizing the funnel to fix a tracking break | Weeks of strategy work aimed at a broken UTM or a consent-banner change | Verify tracking integrity before diagnosing strategy (diagnose.md) |
| Discounting to hit a quarter | A 20% discount at 50% margin needs 67% more units to break even — volume almost never arrives, and the price becomes the new expectation | Time-boxed, reason-coded offers only (pricing.md) |
| Hiring a VP to find the channel | Senior marketers scale motions; they rarely discover one from zero, and the search burns two quarters of salary | Founder or a generalist doer finds the channel; hire the specialist for the channel that already works (operations.md) |
| Launching everything at Tier 1 | Every launch treated as a company moment exhausts the list, the press, and the team; the real launch gets no lift | Tier launches and spend the audience deliberately (launch.md) |
| Buying attribution software to settle an argument | The tool inherits the disagreement and adds a subscription; it cannot see word of mouth or dark social | One holdout test on the disputed line (measurement.md) |
Where Experts Disagree
- Attribution: touch-based software vs. self-reported vs. media mix modeling. Every method lies in a known direction — run them side by side (
measurement.md) rather than picking a church. - Gated vs. ungated content: contacts now vs. reach and trust now, pipeline later. Boundary rule in
content.md. - Category creation vs. competing in an existing category: creation is a multi-year, funding-heavy bet that mostly fails and occasionally wins big. Default to positioning inside an existing category until you hold monopoly-grade differentiation.
- Brand spend in a downturn: cut it to protect runway, or hold share of voice while competitors go quiet (the IPA position — excess SOV is cheapest when others stop bidding). The deciding variable is runway, not conviction: under 12 months, activation wins; above, holding SOV is the higher-return bet.
- Marketing owning a revenue number: shared accountability aligns teams, or it makes marketing hostage to a sales capacity it does not control. Split the difference — marketing owns pipeline created and its conversion to opportunity; sales owns opportunity to close.
Related Skills
More Clawic skills, get them at https://clawic.com/skills/cmo (install if the user confirms):
ceo— executive strategy and board managementpositioning— the positioning statement and competitive frame in depthcro— conversion-rate optimization and landing-page experimentscoo— operations and scaling executionbusiness— strategy validation and planning
Feedback
- If useful, star it: https://clawic.com/skills/cmo
- Latest version: https://clawic.com/skills/cmo
Part of Clawic, the verified skill library. Get this skill: https://clawic.com/skills/cmo.