Agent-Skills.md

Agent Skills: Marketing Analytics

Measure, report on, and optimize marketing performance across channels and campaigns. Trigger on requests for performance dashboards, campaign result analysis, channel metrics review (email, social, paid ads, SEO, content), trend analysis, forecasting, attribution modeling, A/B testing guidance, or actionable optimization recommendations.

UncategorizedID: vm0-ai/vm0-skills/marketing-analytics

Repository

vm0-ai/vm0-skills
vm0-ai
508

Install this agent skill to your local

pnpm dlx add-skill https://github.com/vm0-ai/vm0-skills/tree/HEAD/marketing-analytics

Skill Files

Browse the full folder contents for marketing-analytics.

Download Skill

Loading file tree…

marketing-analytics/SKILL.md

Skill Metadata

Name
marketing-analytics
Description
Measure, report on, and optimize marketing performance across channels and campaigns. Trigger on requests for performance dashboards, campaign result analysis, channel metrics review (email, social, paid ads, SEO, content), trend analysis, forecasting, attribution modeling, A/B testing guidance, or actionable optimization recommendations.

Marketing Analytics

Frameworks for tracking, interpreting, and acting on marketing data across channels, campaigns, and the full customer funnel.

Channel Metric Reference

Email

| Metric | How It Is Calculated | Typical Range | Insight Provided | |--------|---------------------|---------------|-----------------| | Delivery rate | Delivered / Sent | 95-99% | Sender reputation and list hygiene | | Open rate | Unique opens / Delivered | 15-30% | Subject line and sender name effectiveness | | Click-through rate (CTR) | Unique clicks / Delivered | 2-5% | Relevance of content and CTA | | Click-to-open rate (CTOR) | Unique clicks / Unique opens | 10-20% | In-email content quality among openers | | Unsubscribe rate | Unsubscribes / Delivered | Below 0.5% | Audience-content fit and send frequency tolerance | | Bounce rate | Bounces / Sent | Below 2% | List data quality | | Conversion rate | Conversions / Delivered | 1-5% | Full-funnel email performance | | Revenue per send | Total revenue / Emails sent | Varies | Direct monetary contribution | | List growth rate | (New subs - Unsubs) / Total list | 2-5% per month | Audience acquisition health |

Social Platforms

| Metric | How It Is Calculated | Insight Provided | |--------|---------------------|-----------------| | Impressions | Times content appeared in feeds | Distribution breadth | | Reach | Unique users who saw content | Audience coverage | | Engagement rate | (Reactions + Comments + Shares) / Reach | Content resonance | | Click-through rate | Link clicks / Impressions | Ability to drive traffic | | Follower growth rate | Net new followers / Total followers per period | Audience expansion pace | | Share/Repost rate | Shares / Reach | Virality and advocacy signal | | Video view rate | Views / Impressions | Hook effectiveness for video | | Video completion rate | Completed views / Total views | Content quality and length fit | | Share of voice | Your mentions / Category total mentions | Competitive visibility |

Paid Advertising (Search and Social)

| Metric | How It Is Calculated | Insight Provided | |--------|---------------------|-----------------| | Impressions | Times the ad appeared | Budget utilization and audience sizing | | Click-through rate (CTR) | Clicks / Impressions | Creative and targeting relevance | | Cost per click (CPC) | Spend / Clicks | Traffic generation efficiency | | Cost per thousand impressions (CPM) | Spend per 1,000 impressions | Awareness cost efficiency | | Conversion rate | Conversions / Clicks | Landing page and offer effectiveness | | Cost per acquisition (CPA) | Spend / Conversions | Full-funnel cost efficiency | | Return on ad spend (ROAS) | Revenue / Ad spend | Revenue generation return | | Quality Score (search) | Platform relevance rating (1-10) | Alignment of ad, keyword, and destination | | Frequency | Average exposures per user | Ad fatigue risk indicator | | View-through conversions | Conversions from users who saw but did not click | Influence of display and awareness placements |

Organic Search / SEO

| Metric | How It Is Calculated | Insight Provided | |--------|---------------------|-----------------| | Organic sessions | Visits originating from search engines | Overall SEO health | | Keyword positions | Rank for target search terms | Search result visibility | | Organic CTR | Clicks / Search impressions | Title and meta description appeal | | Indexed pages | Pages present in the search index | Crawlability and site architecture | | Domain authority | Third-party composite score | Aggregate site strength | | Backlink count | External domains linking inward | Off-page authority and content value | | Page speed | Time to interactive | UX quality and ranking signal | | Organic conversion rate | Conversions / Organic sessions | Intent alignment and content quality | | Top organic entry pages | Most-visited pages from search | Highest-performing SEO content |

Content Performance

| Metric | How It Is Calculated | Insight Provided | |--------|---------------------|-----------------| | Pageviews | Total views across content pages | Content reach | | Unique visitors | Distinct users consuming content | True audience size | | Average time on page | Duration spent on content pages | Depth of engagement | | Bounce rate | Single-page sessions / All sessions | Content-audience alignment and UX | | Scroll depth | Percentage of page scrolled | Engagement persistence | | Social shares | Times content was distributed socially | Audience advocacy | | Backlinks generated | External links earned by content | SEO value and authority | | Leads attributed | Leads traced to content interaction | Conversion power | | Content ROI | Attributed revenue / Production cost | Investment return |

Pipeline and Revenue Metrics

| Metric | How It Is Calculated | Insight Provided | |--------|---------------------|-----------------| | Marketing qualified leads (MQLs) | Leads passing marketing qualification criteria | Top-of-funnel output | | Sales qualified leads (SQLs) | MQLs accepted by the sales team | Lead quality | | MQL-to-SQL conversion | SQLs / MQLs | Marketing-sales alignment | | Pipeline created | Dollar value of new opportunities | Marketing revenue impact | | Pipeline velocity | Speed of deal progression | Campaign urgency and quality signal | | Customer acquisition cost (CAC) | Total marketing + sales spend / New customers | Acquisition efficiency | | CAC payback period | Months to recoup CAC from revenue | Unit economics viability | | Marketing-sourced revenue | Revenue from marketing-originated deals | Direct marketing contribution | | Marketing-influenced revenue | Revenue from deals with any marketing touchpoint | Broader marketing footprint |

Report Structures

Weekly Snapshot

Designed for rapid team consumption:

  • Three headline metrics with week-over-week movement
  • Wins: 1-2 data-backed highlights
  • Watch items: 1-2 areas requiring attention with supporting numbers
  • Upcoming actions: 3-5 priorities for the week ahead

Monthly Performance Review

Standard format for stakeholder reporting:

  1. Executive summary (3-5 sentences)
  2. Core metrics table with month-over-month and target comparisons
  3. Channel-level performance breakdown
  4. Campaign results and highlights
  5. What succeeded and what underperformed, with working hypotheses
  6. Recommendations and priorities for the coming month
  7. Budget spent vs. planned

Quarterly Strategic Review

For leadership-level analysis:

  1. Quarter results against stated goals
  2. Year-to-date progress and trajectory
  3. Channel-by-channel ROI assessment
  4. Campaign portfolio performance summary
  5. Competitive and market landscape observations
  6. Strategic recommendations for the next quarter
  7. Budget proposal and reallocation plan
  8. Experiment outcomes and key learnings

Dashboard Construction Principles

  • Feature the metrics that tie directly to business goals, not vanity numbers
  • Display trends over multiple periods rather than isolated data points
  • Provide comparison anchors: prior period, target, industry benchmark
  • Apply uniform color signaling: green for on-track, yellow for at-risk, red for off-track
  • Organize by funnel stage or the business question being answered
  • Confine the dashboard to a single screen; relegate granular data to an appendix
  • Match the refresh cadence to the decision cadence (real-time for paid media, weekly for content)

Trend Analysis and Projection

Spotting Patterns

When examining performance data, investigate:

  1. Sustained direction: is the metric consistently rising, falling, or flat across 4+ consecutive periods?
  2. Turning points: at what moment did the trajectory change, and what event coincided?
  3. Cyclical patterns: are there recurring fluctuations by day of week, month, or quarter?
  4. Outliers: isolated spikes or dips — what triggered them, and could the cause be replicated or avoided?
  5. Predictive signals: which metrics shift first and foreshadow downstream outcomes?

Analytical Process

  1. Plot the metric across time with at least 8-12 data points for statistical relevance
  2. Characterize the overall trajectory (rising, declining, stable, or oscillating)
  3. Quantify the rate of change — is the trend accelerating or flattening?
  4. Layer in external events (campaign launches, product updates, market shifts)
  5. Benchmark against targets or industry norms
  6. Look for correlations with related metrics
  7. Formulate causal hypotheses and design experiments to test them

Projection Techniques

  • Trend extension: project the existing trajectory forward (works best for stable metrics)
  • Rolling average: average the most recent 3-6 periods to dampen noise
  • Year-over-year overlay: use the prior year's seasonal pattern, adjusted for a growth coefficient
  • Funnel arithmetic: forecast outputs from inputs (X leads at Y% conversion rate yields Z customers)
  • Scenario planning: model optimistic, expected, and pessimistic cases

Projection Guardrails

  • Near-term forecasts (1-3 months) carry far more reliability than long-range ones
  • Projections built on fewer than 12 data points should be labeled low-confidence
  • External disruptions (market shifts, competitive moves, economic changes) can invalidate trend-based models
  • Always express forecasts as ranges rather than single numbers

Attribution Fundamentals

Why Attribution Matters

Buyers rarely convert after a single interaction. Attribution assigns credit across the multiple touchpoints that precede a conversion, informing channel investment decisions.

Standard Attribution Models

| Model | Mechanism | Strength | Weakness | |-------|----------|----------|----------| | Last interaction | All credit to the final touchpoint | Identifies closing channels | Overlooks awareness and nurture | | First interaction | All credit to the initial touchpoint | Highlights discovery channels | Ignores conversion drivers | | Even distribution | Equal credit across all touchpoints | Acknowledges every channel | Fails to reflect relative influence | | Recency-weighted | Increasing credit as touchpoints approach conversion | Balances awareness and closing | Can undervalue early awareness | | Position-based (40/20/40) | Heavy credit to first and last, remainder split across the middle | Honors both discovery and conversion | Somewhat arbitrary weight assignment | | Algorithmic | Machine-learned credit based on conversion path data | Most reflective of actual influence | Demands large conversion volumes |

Practical Attribution Advice

  • If you have no attribution system, begin with last-interaction — it is the simplest and most immediately actionable
  • Contrast first-interaction and last-interaction views to learn which channels drive discovery vs. closure
  • Position-based (40/20/40) is a pragmatic default for most B2B organizations
  • Algorithmic models need high conversion volumes to produce statistically sound results
  • Treat attribution as directional intelligence, never as absolute truth
  • Any multi-touch model is more informative than a single-touch model, and any model outperforms none

Attribution Traps

  • Optimizing a single channel based on single-touch data can starve the rest of the funnel
  • Awareness-oriented channels (display, organic social, PR) will consistently underperform in last-touch reports
  • Conversion-oriented channels (branded search, retargeting) will consistently underperform in first-touch reports
  • Self-reported attribution ("How did you hear about us?") offers useful qualitative signal but is unreliable for quantitative allocation
  • Cross-device and cross-channel tracking gaps guarantee that attribution data is always incomplete

Optimization Methodology

Systematic Improvement Process

  1. Detect: which metrics fall short of targets or benchmarks?
  2. Locate: where in the funnel does the breakdown occur? (impressions, clicks, conversions, retention)
  3. Theorize: what is causing the shortfall? (targeting, messaging, creative, offer design, timing, technical issues)
  4. Rank: which interventions promise the greatest impact relative to effort?
  5. Experiment: run a controlled test to validate or disprove the hypothesis
  6. Evaluate: did the metric improve meaningfully?
  7. Act: scale successful changes broadly; iterate on inconclusive or negative results

Intervention Levers by Funnel Position

| Funnel Position | Warning Sign | Available Levers | |----------------|-------------|-----------------| | Awareness | Low impressions, limited reach | Budget levels, targeting parameters, channel mix, ad format | | Interest | Low CTR, weak engagement | Creative execution, headline copy, content hooks, audience refinement | | Consideration | High bounce rate, low dwell time | Page content, load speed, relevance alignment, user experience | | Conversion | Low conversion rate | Offer structure, CTA wording, form complexity, trust elements, page layout | | Retention | Elevated churn, declining re-engagement | Onboarding flow, email sequences, product experience, support quality |

Impact-Effort Prioritization

Score every optimization idea on two axes:

Impact (potential metric movement):

  • High: directly addresses the primary bottleneck
  • Medium: improves a contributing factor
  • Low: yields incremental gains

Effort (implementation difficulty):

  • Low: copy tweak, targeting adjustment, quick A/B test
  • Medium: new creative asset, page redesign, workflow modification
  • High: new tooling, cross-team initiative, major content production

Execution order:

  1. High impact, low effort — execute immediately
  2. High impact, high effort — plan and staff
  3. Low impact, low effort — pursue if bandwidth allows
  4. Low impact, high effort — defer or deprioritize

Experimentation Discipline

  • Isolate a single variable per test for interpretable results
  • Lock in the success metric before the test begins
  • Calculate the required sample size in advance and resist ending tests prematurely
  • Run each test for at least one complete business cycle (usually a full week for B2B)
  • Record all experiments and outcomes, including negative and null results
  • Circulate learnings across the team — a test that confirms the current approach still builds confidence

Ongoing Optimization Rhythm

  • Daily: check paid campaign pacing, flag anomalies, review ad approval status
  • Weekly: assess channel-level performance, pause lagging efforts, amplify winners
  • Biweekly: rotate ad creative and launch new test variants
  • Monthly: conduct a comprehensive performance review, surface new optimization opportunities, refresh projections
  • Quarterly: reassess channel strategy, budget distribution, and audience targeting at a strategic level