К основному содержанию
скилл CX Codex

marketing-analytics — скилл для Claude Code: аналитика и установка

Скилл marketing-analytics для Claude Code разбирает маркетинговые данные и помогает планировать оптимизацию кампаний. Запускается на запросы об A/B-тестах, конверсиях, исследованиях клиентов и аналитических отчётах.

MIT ★ 43 обновлён 2026-11-16

Установить в Claude Code · Через Skills CLI (Vercel) Все способы и платформы ↓
npx skills add shalintripathi/saas-marketing-agents --skill marketing-analytics -a claude-code
Открыть на GitHub

Что делает marketing-analytics

Скилл направляет запрос к специалисту по нужной задаче: анализу эффективности кампаний, оптимизации конверсии, исследованию клиентов или рассказу о данных. Он охватывает также построение маркетинговых дашбордов, моделей атрибуции и анализ воронки. Перед работой ищет brand-context.md в корне проекта, ./.claude/brand-context.md и ./docs/brand-context.md.

Если файл найден, скилл передаёт его правила подключённым специалистам. Если файла нет, он сообщает об этом и предлагает шаблон templates/brand-context.md или подготовить черновик после интервью либо изучения материалов компании. Допущения нужно помечать явно; названия клиентов, метрики и результаты нельзя выдумывать. Без подтверждающих сведений скилл добавляет маркер [NEEDS INPUT: …].

Какие задачи решает

  • разобрать эффективность маркетинговых кампаний
  • найти узкие места воронки и подготовить CRO-тест
  • спланировать исследование клиентов или опрос
  • собрать маркетинговые выводы в дашборд или рассказ о данных

Примеры запросов агенту

  • Разбери эффективность наших маркетинговых кампаний и покажи основные факторы.
  • Найди узкие места воронки и предложи план A/B-теста.
  • Составь план интервью с клиентами и обозначь ограничения выборки.

Как установить marketing-analytics

1Через Skills CLI (Vercel)

npx skills add shalintripathi/saas-marketing-agents --skill marketing-analytics -a claude-code

2Вручную: положить файлы

~/.claude/skills/marketing-analytics/SKILL.md

Папка скилла целиком; для одного проекта — .claude/skills/marketing-analytics/

По шагам, с командами для каждой платформы: Как установить скилл в Claude Code: пять способов

Вопросы о marketing-analytics

Как установить marketing-analytics в Claude Code?

Выполни npx skills add shalintripathi/saas-marketing-agents --skill marketing-analytics -a claude-code. Файл скилла находится по пути ~/.claude/skills/marketing-analytics/SKILL.md.

Что делает marketing-analytics?

Скилл маршрутизирует задачи по маркетинговой аналитике к профильным специалистам: от разбора кампаний и воронки до CRO-тестов и исследований клиентов. Также помогает готовить дашборды, модели атрибуции и объяснения данных.

Зачем marketing-analytics нужен файл brand-context.md?

В нём хранятся сведения о целевой аудитории, позиционировании, голосе бренда и ограничениях. Скилл ищет файл в корне проекта, ./.claude/brand-context.md и ./docs/brand-context.md; если файла нет, предлагает шаблон или подготовить черновик.

Исходный текст

SKILL.md · MIT
---
name: marketing-analytics
description: "Marketing analytics and data-driven optimization for B2B SaaS campaigns. Use this skill when analyzing marketing performance, optimizing conversion rates, running A/B tests, conducting customer research, creating marketing dashboards, building attribution models, analyzing funnel metrics, or storytelling with data. Also triggers on: marketing ops, performance analytics, CRO, A/B testing, data storytelling, customer insights, tracking plan, event taxonomy, UTM taxonomy, GA4 audit, exclude internal traffic from GA4, our own team inflates the conversion numbers, GA4 AI Assistant channel, how do I track ChatGPT traffic in GA4, Source Group, staging or test traffic in analytics, unknown hostname in GA4, GA4 data filters, hostname allowlist, consent mode, instrumentation contract, sample size calculation, minimum detectable effect, test duration, not enough traffic to A/B test, testing a pop-up or chat widget only some visitors see, triggered analysis, statistical significance, Bayesian vs frequentist, sample ratio mismatch, experiment feasibility, martech stack, marketing tech stack audit, tool consolidation, build vs buy, vendor evaluation, shadow IT tools, ESP/CRM/platform migration, switching cost, cutover plan, suppression list migration, data portability, tool adoption after go-live, nobody uses the new tool, we bought it but the team still uses spreadsheets, shelfware prevention, customer interviews, survey design, sampling frame, non-response bias, how many people should I interview, thematic saturation, voice of customer, jobs-to-be-done, buyer persona research, is this finding real, data enrichment, which enrichment provider should we buy, waterfall enrichment, contact data quality, match rate, our list data is bad, email verification, catch-all domain, accept-all, bounce rate from bad data, B2B data decay, how often should we re-verify contacts, ZoomInfo vs Apollo vs Clearbit evaluation, cost per usable record, where did this list come from, GDPR Article 14 notice, purchased list compliance, suppression screening, TAM list building, demand planning, marketing plan, how many leads do we need, what pipeline target should marketing carry, funnel model, top-down vs bottom-up plan, plan of record, annual marketing planning, quarterly planning, pipeline coverage we have to create, sales cycle lag, our leads cannot close this quarter, carry-in pipeline, scenario planning, conservative moderate aggressive forecast, stretch target, reforecast, re-forecast triggers, assumption register, capacity ceiling, addressable audience ceiling, marginal cost per lead, blended CPL is wrong, we missed plan and nobody knows why, AI policy for the marketing team, AI governance, AI acceptable use policy, which AI tools are we allowed to use, can I paste customer data into ChatGPT, shadow AI, AI use register, human review of AI output, when can we stop reviewing every AI draft, can the agent send on its own, the AI made a mistake now what, pull back AI autonomy, AI evals, our AI vendor changed the model, AI output went out wrong, prompt injection, can our AI read prospect replies safely, lethal trifecta, the AI leaked internal notes, turn off auto-generated ad assets, auto-applied recommendations, AI literacy training, EU AI Act Article 4, duplicate records in the CRM, dedupe before import, which value wins in a merge, CRM cleanup."
---

# Marketing Analytics

## Step 0 (always first): Load brand context

**Before producing any deliverable, look for a `brand-context.md` file** in the user's project root (also check `./.claude/brand-context.md` and `./docs/brand-context.md`). It holds the company's ICP, positioning, messaging pillars, citable proof, voice, banned words, and compliance constraints.

- **If it exists:** read it in full and treat it as binding for this run. Hand its contents to every specialist agent you route work to, alongside the task brief. Its "Rules for agents reading this file" section overrides an agent's own defaults.
- **If it does not exist:** say so, point the user at the template ([`templates/brand-context.md`](../../templates/brand-context.md)), and offer to generate a filled draft by interviewing them or by reading their website and existing content. Then proceed with explicitly-labelled assumptions — never silently invented ones.

**Non-negotiable regardless of which path applies:** do not invent customer names, metrics, funding, integrations, certifications, or outcomes. Only proof recorded in `brand-context.md` (or supplied directly in the request) may be used as fact. Where a claim would help but no evidence exists, emit a `[NEEDS INPUT: …]` marker in the deliverable rather than a plausible-sounding guess.

---

## What This Is

Marketing Analytics brings together performance analysts, conversion rate optimization specialists, customer insights researchers, data storytellers, and marketing operations architects to transform data into decisions. This skill orchestrates your analytics practice to measure what matters (pipeline impact, customer quality, efficient CAC), identify optimization opportunities (conversion rate improvements, targeting precision, channel efficiency), and communicate findings to stakeholders. Whether you're building a marketing dashboard, running an A/B test program, conducting customer research to understand buyer behavior, or presenting performance analysis to leadership, Marketing Analytics routes your request to the right specialist and ensures your data drives better marketing decisions.

## The Team: 9 Specialist Agents

| # | Agent | File | What They Do |
|---|-------|------|-------------|
| 1 | Performance Analyst | `agents/analytics-performance-analyst.md` | Measures and optimizes marketing performance across channels and campaigns, tracks KPIs, and identifies performance drivers through data analysis and reporting |
| 2 | Conversion Rate Optimizer | `agents/analytics-conversion-rate-optimizer.md` | Designs and executes CRO testing programs, analyzes conversion funnels, identifies friction points, optimizes user experience for higher conversion rates, and gates every experiment on whether the funnel can actually power it (sample size, duration, comparability horizon), including a triggered read when only part of the traffic ever sees the change |
| 3 | Customer Insights Researcher | `agents/analytics-customer-insights-researcher.md` | Designs and runs customer research (interviews, surveys, VoC mining) and grades what it can claim — sampling frame and non-response, evidence levels, cross-channel corroboration, saturation, jobs-to-be-done and research-based personas |
| 4 | Data Storyteller | `agents/analytics-data-storyteller.md` | Translates data analysis into clear narratives and compelling visualizations that communicate findings and drive decision-making |
| 5 | Marketing Ops Architect | `agents/analytics-marketing-ops-architect.md` | Designs the data architecture inside the stack — MAP/CRM integration, lead lifecycle and scoring, field governance, duplicate prevention (resolve before create) and merge survivorship rules — authors the marketing instrumentation contract (tracking plan, event taxonomy, UTM taxonomy) that every measurement audit grades against, audits GA4 down to which traffic is let in (internal, developer and hostname data filters, sized in Testing before they are switched on) and how the native AI Assistant channel and Source Group reconcile against your own AI referrer list, implements tracking and attribution, and builds dashboards and reporting systems that enable data-driven marketing |
| 6 | GTM Data Strategist | `agents/analytics-gtm-data-strategist.md` | Owns the go-to-market data that arrives from outside — the data requirement written before shopping, provider bake-offs on a held-out sample of your own accounts, coverage scored separately from accuracy and wrong answers, identity-match corroboration, enrichment waterfalls ordered by measured cost per usable record, a decay rate measured rather than quoted, verification states including accept-all, suppression screening, and the provenance and lawful-basis record (GDPR Article 14) behind every acquired list |
| 7 | MarTech Stack Strategist | `agents/analytics-martech-stack-strategist.md` | Owns the martech stack as a portfolio: capability-first tool selection and build-vs-buy, shadow-tool and overlap rationalization, and the true switching cost of replacing a system — including the suppression record, engagement history, personalization layer and measurement continuity a platform migration puts at risk — and proving after go-live, from the system's own records, that the team actually adopted it |
| 8 | Demand Planner | `agents/analytics-demand-planner.md` | Builds the number before the period starts — the revenue target decomposed backwards through the funnel into stage volumes, an assumption register where every conversion rate and cycle time is a dated own-funnel measurement or is stamped inferred, the sales-cycle lag that decides which period a lead can land in (carry-in, carry-out, latest useful creation date), the sales / production / addressable-audience ceilings that cap what a plan may promise, marginal rather than blended cost per source, a scenario band derived from your own plan-versus-actual dispersion instead of a borrowed ±20%, declared re-forecast triggers, and a versioned plan of record that makes a miss diagnosable |
| 9 | Marketing AI Operations Lead | `agents/analytics-marketing-ai-ops-lead.md` | Governs how the team uses AI — an AI use register per use case, legalize-don't-ban intake for shadow AI, data-class rules for which tool may receive what, an autonomy ladder with a never-generate list and step-down triggers that fire without approval, a prompt-injection check (private data + untrusted content + outbound path caps a workflow at human review), eval sets re-run on every model/prompt/vendor change, default-on platform AI settings, kill switches, an AI incident log, and an EU AI Act Art. 4 literacy record by role |

## How to Use

### Routing User Requests

**Marketing Performance Analysis** → Performance Analyst
- "Analyze the performance of our latest demand generation campaign"
- "Create a monthly marketing dashboard showing KPIs by channel"
- "Identify which channels are driving the most qualified pipeline"
- "Compare performance across campaigns to identify what's working"

**Demand & Pipeline Planning** → Demand Planner
- "How many leads do we need next year to hit the revenue target?"
- "Build the bottom-up marketing plan and reconcile it against the number finance gave us"
- "Our Q1 plan assumes leads created in January close in March — is that possible?"
- "Give me a conservative / planned / upside band with the triggers that tell us which one we are in"
- "We missed plan — which assumption was wrong?"

**Conversion Rate Optimization** → Conversion Rate Optimizer
- "Design a CRO testing program for our landing pages"
- "Analyze our conversion funnel and identify where prospects drop off"
- "Run A/B test comparing headline variations on our homepage"
- "Optimize form fields to increase lead capture completion rate"
- "How much traffic do we need to A/B test this page, and how long will it take?"
- "We don't have enough traffic to reach significance — what should we do instead?"
- "Should we use a Bayesian test because our sample is small?"

**Customer Research & Insights** → Customer Insights Researcher
- "Conduct survey research to understand what messages resonate with our target buyer"
- "Interview customers to understand decision process and evaluation criteria"
- "Analyze user behavior data to identify content preferences"
- "Research competitive buyer perception to understand market positioning opportunity"

**Data Communication & Visualization** → Data Storyteller
- "Create a visual dashboard showing marketing funnel metrics"
- "Present Q3 performance analysis to executive leadership"
- "Design infographic summarizing campaign results for stakeholders"
- "Build a customer journey visualization based on analytics data"

**Marketing Systems & Infrastructure** → Marketing Ops Architect
- "Design our marketing attribution model (multi-touch vs. first-touch vs. last-touch)"
- "Build a marketing dashboard infrastructure with real-time data updates"
- "Implement tracking and data collection for marketing funnel"
- "Create a data dictionary and standardized metrics definitions for marketing"

**MarTech Stack, Tool Selection & Platform Migration** → MarTech Stack Strategist
- "Audit our martech stack — what are we paying for that we don't use?"
- "We have too many tools and they don't talk to each other"
- "Should we buy this tool, build it, or does something we already own do it?"
- "What does it actually cost us to move off our current ESP / MAP / CRM?"
- "Plan the cutover: what has to carry across, and what breaks if it doesn't?"
- "How do we make sure our unsubscribes survive the migration?"
- "We rolled out the new MAP three months ago and the team still builds lists in spreadsheets"

**Marketing AI Governance & AI Operations** → Marketing AI Operations Lead
- "We need an AI policy for the marketing team — what's allowed?"
- "Can I paste our customer list / call transcripts into ChatGPT?"
- "Half the team uses AI tools we never approved — what do we do?"
- "Our AI email personalization is about to send on its own — is it ready?"
- "The vendor changed the model and the outputs changed — how would we even know?"
- "Which AI features are switched on by default in our ad accounts and ESP?"
- "What AI-literacy training do we need to show under the EU AI Act?"

### Execution Model

**Phase 1: Baseline & Measurement (1-2 weeks)**
- Marketing Ops Architect establishes current measurement infrastructure: what's being tracked, what's missing, data quality issues
- Performance Analyst establishes baseline metrics: current funnel conversion rate, cost per lead, cost per acquisition, channel performance
- Identify measurement gaps: what's not being tracked that should be (first-touch vs. last-touch, buyer quality signals, sales pipeline influence)
- Define KPIs and success metrics for next phase of work

**Phase 2: Analysis & Diagnosis (2-3 weeks)**
- Performance Analyst analyzes historical performance data to identify trends, patterns, anomalies
- Customer Insights Researcher conducts qualitative research (surveys, interviews) to understand why metrics look the way they do
- Conversion Rate Optimizer identifies conversion funnel bottlenecks: where do prospects drop off most
- Develop hypothesis for optimization opportunities
- Data Storyteller synthesizes findings into narrative and visualizations

**Phase 3: Experimentation & Testing (4-12 weeks depending on scope)**
- Conversion Rate Optimizer designs and launches A/B testing program with prioritized hypotheses
- Marketing Ops Architect ensures testing infrastructure (statistical significance tracking, proper audience segmentation, variant tracking)
- Customer Insights Researcher collects feedback on tested variations (surveys, interviews to understand why winners won)
- Performance Analyst monitors test results weekly, identifies winners, and tracks cumulative impact
- Data Storyteller communicates test results and learnings to stakeholders

**Phase 4: Optimization & Scaling (ongoing)**
- Scale winning variations (increase traffic to better-performing page, increase spend on better-performing channel)
- Iterate next set of tests based on learning from prior tests
- Customer Insights Researcher continuously monitors sentiment and satisfaction (NPS, CSAT, customer interviews)
- Performance Analyst tracks improvement impact on downstream metrics (pipeline quality, customer lifetime value)
- Marketing Ops Architect maintains data freshness and dashboard accuracy

### Testing Methodology

**CRO Testing Framework**
1. Hypothesis formation (based on data analysis or customer research insight)
   - Expected: Landing page headline focusing on outcome (vs. feature) will increase conversion rate
   - Why: Customer research showed buyers care about time savings more than features
   - Metric: Track conversion rate, form completion rate, downstream pipeline quality

2. Test design (control, test variant, sample size, duration)
   - Control: current page (baseline performance)
   - Variant: new headline focusing on outcome
   - Sample size: 100 conversions per variant to reach 95% statistical significance
   - Duration: 2 weeks or until sample size achieved

3. Execution (launch test, monitor for data quality)
   - Route 50% traffic to control, 50% to variant
   - Ensure proper tracking implementation (no missing conversions, no duplicate counts)
   - Monitor daily for anomalies (traffic drop, conversion spike indicating technical error)

4. Analysis (determine winner, measure impact, document learning)
   - Did test win or lose (statistical significance threshold: 95% confidence, 5% significance level)
   - If winner: magnitude of improvement (e.g., 12% lift in conversion rate = 0.5% absolute improvement)
   - If loser: why did it lose? Did it hurt downstream metrics (lead quality, sales conversion)?
   - Document learning: which element drove result (headline, color, form field, CTA placement)

5. Iteration (scale winner, test next hypothesis)
   - Scale winning variant to 100% traffic (or keep testing if hold-back analysis desired)
   - Identify next testing hypothesis based on remaining friction points
   - Queue up next test for launch

**Testing Prioritization**
- Prioritize tests by estimated impact × confidence level
- High-traffic pages (more statistical power) → test bigger changes
- Low-traffic pages → test smaller changes (less power) or focus on traffic growth
- Test single variable per test (isolates learnings)
- Test highest-impact elements first: traffic source → landing page → form completion → downstream quality

**Statistical Rigor**
- Sample size calculation: depends on baseline conversion rate, desired lift, desired significance level
- Significance threshold: 95% confidence (5% alpha, meaning 5% chance of false positive)
- Running time: continue test until sample size achieved or 2-4 weeks, whichever comes first
- Avoid peeking: don't declare winner until test complete (peeking inflates error rate)

### Metrics Hierarchy

**Top-Level KPIs (North Star Metrics)**
- Marketing-Influenced Pipeline: revenue impact of marketing activities (pipeline value marketing generated)
- Marketing Qualified Leads (MQLs): leads that meet quality threshold for sales handoff
- Customer Acquisition Cost (CAC): fully-loaded cost to acquire new customer (including marketing + sales)
- Customer Lifetime Value (CLV): expected revenue from customer over lifetime

**Mid-Level Metrics (Channel Performance)**
- Cost Per Lead (CPL): average marketing spend to acquire one lead
- Lead-to-Opportunity Conversion: percentage of leads that convert to sales-accepted opportunity
- Opportunity-to-Customer Conversion: percentage of opportunities that close to customer
- Cost Per Opportunity: marketing spend to generate one sales-accepted opportunity
- Cost Per Acquisition: marketing spend to close one customer deal

**Funnel Metrics (Conversion Analysis)**
- Website Traffic: monthly unique visitors by source (organic, paid, direct, referral)
- Page Conversion Rate: percentage of page visitors who take desired action (form submission, email signup)
- Form Completion Rate: percentage of form starters who complete form submission
- Demo/Trial Signup Rate: percentage of engaged prospects who request demo or free trial
- Sales Accepted Lead Rate: percentage of inbound leads meeting sales qualification criteria

**Campaign Metrics (Campaign Performance)**
- Click-Through Rate (CTR): percentage of ad impressions that receive clicks (LinkedIn: 0.5-2%, Google: varies by audience)
- Cost Per Click (CPC): average cost per paid click (compare to historical benchmark and competitor spend)
- Video Completion Rate: percentage of video viewers who watch 75%+ of video (YouTube 25-40%, in-feed 15-25%)
- Lead Quality Score: MQL to SAL conversion rate (higher = better quality leads from source)
- Return on Ad Spend (ROAS): revenue generated per dollar of paid ad spend

**Customer Quality Metrics**
- Sales Cycle Length: average time from opportunity creation to closed customer
- Win Rate: percentage of opportunities that close to customer (competitive benchmark varies)
- Average Contract Value: average customer deal size
- Customer Retention Rate: percentage of customers retained year-over-year
- Net Revenue Retention: revenue retention + expansion revenue from existing customers

### Routing by Analytics Use Case

**Landing Page Optimization**
- Conversion Rate Optimizer leads CRO program with testing roadmap
- Marketing Ops Architect ensures tracking and statistical testing infrastructure
- Data Storyteller visualizes funnel and testing results
- Performance Analyst measures impact on downstream metrics (pipeline, customer quality)
- Success metrics: 20-30% conversion rate improvement over 6 months, improved lead quality

**Demand Generation Campaign Analysis**
- Performance Analyst analyzes campaign performance by channel and content
- Customer Insights Researcher surveys engaged prospects to understand preferences
- Conversion Rate Optimizer identifies highest-converting assets and channels
- Data Storyteller visualizes campaign funnel and highlights top performers
- Marketing Ops Architect ensures proper attribution and tracking
- Success metrics: hit campaign pipeline goal, exceed average cost-per-lead benchmark, high lead quality

**Attribution & Analytics Infrastructure**
- Marketing Ops Architect designs multi-touch attribution model
- Performance Analyst implements tracking and validates data quality
- Data Storyteller creates dashboards showing attribution by channel, campaign, source
- Marketing Ops Architect establishes data governance and standardized metrics
- Success metrics: 95%+ tracking accuracy, real-time dashboard updates, executive confidence in data

**Customer Research Program**
- Customer Insights Researcher conducts qualitative and quantitative research
- Surveys: 100-200 target buyers on messaging, positioning, feature priorities
- Interviews: 10-20 deep-dive conversations with buyers and existing customers
- Behavioral analysis: analyze user behavior on website, demo usage, content preferences
- Data Storyteller synthesizes findings into narrative and visualizations
- Success metrics: identify 5-10 actionable customer insights, 3+ changes implemented based on research

## Output Standards

### Analytics Quality Checklist

**Data Accuracy**
- Validation: Data validated against source systems (CRM, ad platform, GA) for accuracy
- Completeness: 95%+ of expected events tracked and recorded
- Freshness: Dashboards updated daily or hourly for real-time decision-making
- Reconciliation: Monthly reconciliation between systems to ensure consistency
- Documentation: Data dictionary documenting every metric, calculation method, data source

**Funnel Analysis**
- All stages defined: awareness → engagement → MQL → SAL → opportunity → customer
- Conversion rates calculated: percentage of visitors reaching each stage
- Benchmarks established: compare conversion rates to historical baseline and industry benchmark
- Segment analysis: conversion rates by traffic source, campaign, buyer persona, vertical
- Bottleneck identification: which stage has lowest conversion or highest drop-off

**Testing Rigor**
- Sample size: adequate sample size (100+ conversions per variant) to reach 95% statistical significance
- Duration: test runs until sample size achieved or 2-4 weeks minimum
- Single variable: each test isolates one change (control constant, one variable changed)
- Statistical significance: 95% confidence level, 5% significance threshold
- Learning documentation: document what we learned and how it applies to future tests

**Dashboard Quality**
- Clarity: metrics labeled clearly, units defined (rates as %, numbers as count or $)
- Timeliness: updated daily or hourly depending on decision-making frequency
- Actionability: dashboard shows what's working/not working, not just data points
- Drilling down: ability to click through to see underlying data (top traffic sources, best-performing content, etc.)
- Comparisons: shows trends over time and comparisons to benchmarks/targets

### Deliverable Specifications

**Marketing Dashboard**
- Executive dashboard (1-2 pages): top KPIs, traffic, pipeline, campaign ROI, quarter performance vs. goal
- Funnel dashboard: conversion rates by stage, segment analysis (by source, campaign, persona), bottleneck identification
- Channel dashboard: performance by channel (organic, paid, direct, referral, email), cost per lead, conversion rate, trend
- Campaign dashboard: active campaigns, performance vs. goal, cost per lead, lead volume, quality metrics
- Testing dashboard: active tests, results, statistical significance, learning captured
- Daily dashboard: refresh rate hourly, daily KPI performance, alert if metrics miss thresholds

**CRO Testing Roadmap**
- Testing hypothesis bank: 20-30 prioritized hypotheses with expected impact and effort estimate
- Hypothesis: description of test variant, expected outcome, business rationale, success metric
- Estimated lift: expected conversion rate improvement based on similar tests or industry benchmarks
- Sample size & duration: how long test will run to reach statistical significance
- Timeline: Quarters 1-4 roadmap with hypothesis sequence
- Resource requirements: design effort, analytics effort, traffic required

**Performance Analysis Report**
- Executive summary: key findings, performance vs. goal, recommended actions
- Performance overview: KPIs, trends, comparison to prior period and target
- Channel analysis: performance by traffic source (organic, paid, direct), cost per lead, conversion rate
- Campaign analysis: performance by campaign, content, topic, top performers
- Customer quality analysis: lead quality by source, sales cycle length, customer retention
- Competitive benchmarking: performance vs. industry benchmarks
- Recommendations: 3-5 prioritized actions to improve performance

**Customer Insights Report**
- Research methodology: survey sample size, interview count, data collection approach
- Key findings: 5-10 major insights about buyer behavior, preferences, decision criteria
- Messaging preferences: what messages resonate, which value props matter most by persona
- Competitive positioning: how buyers perceive competitive alternatives, key decision factors
- Buying process: stages in buyer journey, key decision drivers by stage
- Persona profiles: updated or confirmed persona attributes, priorities, concerns
- Recommendations: how findings should inform marketing strategy, messaging, product

**CRO Test Results Document**
- Test overview: hypothesis, expected outcome, why this test matters
- Test design: control variant, test variant, sample size, duration, audience selection
- Results: conversion rate by variant, percentage lift, statistical significance
- Statistical analysis: confidence level, p-value, sample size achieved
- Impact: downstream metrics (lead quality, sales conversion, customer retention)
- Learning: what we learned, why winner won, how to apply to future tests
- Follow-up: next test to run based on this learning

**Attribution Model Specification**
- Model type: first-touch vs. last-touch vs. multi-touch (time decay, linear, custom)
- Touch definitions: what actions count as "touch" (landing page view, form submission, email click)
- Credit allocation: how credit is allocated among touches (first touch gets 100%, or distributed across journey)
- Window: how long to track touches (7-day, 30-day, 90-day cookie window)
- Implementation: technical specification for tracking and attribution calculation
- Reporting: how results will be reported and updated
- Validation: methodology to validate attribution accuracy

### Quality Review Process

Before publishing any analytics finding or dashboard:

1. **Data Validation**: Is the data accurate and complete? Have we reconciled against source systems?
2. **Statistical Rigor**: Do findings meet statistical significance requirements? Is sample size adequate?
3. **Bias Check**: Could sample bias, selection bias, or confounding variables affect findings?
4. **Clarity**: Can stakeholders understand findings without extended explanation?
5. **Actionability**: Do findings lead to clear recommendations or decisions?
6. **Sensitivity Check**: How sensitive are findings to assumptions? What changes would reverse finding?

### Success Metrics by Analytics Program

**Conversion Rate Optimization**
- Test execution: launch 1-2 tests per week, completion of testing roadmap
- Test quality: 80%+ of tests have adequate sample size and statistical rigor
- Win rate: 40-50% of tests show positive lift (indicates healthy hypothesis generation)
- Conversion improvement: 20-30% cumulative conversion rate improvement over 6 months
- Downstream impact: 10-15% improvement in lead quality or sales cycle length

**Marketing Dashboard & Reporting**
- Dashboard adoption: 90%+ of stakeholders use dashboard for decision-making
- Update frequency: dashboards updated daily or hourly with zero latency delays
- Accuracy: 98%+ data accuracy compared to source systems
- Action tracking: 80%+ of dashboard insights lead to actionable decisions
- Executive confidence: stakeholder survey showing 90%+ confidence in data accuracy

**Customer Research Program**
- Research completion: conduct 20-30 interviews and 100-200 survey responses per quarter
- Insight generation: 5-10 actionable insights per research cycle
- Implementation rate: 80%+ of recommended changes implemented within 90 days
- Impact measurement: track ho...

© shalintripathi · MIT · shalintripathi/saas-marketing-agents

Похожие скиллы

Все скиллы по теме «Данные и базы» →
analytics coreyhaines31 Скилл analytics для Claude Code и Codex помогает спроектировать план трекинга и проверить формулы метрик. Настраивает события в GA4, GTM, PostHog и Mixpanel без искажения когорт. CX Codex +1 97 200установок devrel-analytics Samuel Berthe Скилл devrel-analytics для Claude Code строит план аналитики DevRel: события, источники, идентификаторы и воронки. Связывает посещения документации с регистрациями и объясняет расхождения метрик. CX Codex +1 13 500установок app-analytics appeeky Скилл app-analytics помогает настроить трекинг мобильного приложения, собрать карту событий и оценить воронку покупок. Пригодится при запуске платного трафика или подготовке отчётов в App Store Connect. CX Codex +1 3 600установок analytics-tracking kostja94 Скилл analytics-tracking указывает на актуальные инструкции Open Analytics по событиям GA4, ключевым событиям и User-ID. Нужен при переходе со старого пути скилла. CX Codex +1 2 500установок analytics-and-reporting social-media-skills Скилл analytics-and-reporting превращает сырую статистику соцсетей в практические выводы по методике METER без накрученных метрик вроде лайков или просмотров. CX Codex +1 1 500установок amazon-brand-analytics nexscope-ai Расшифровывает данные Amazon Brand Analytics — Search Frequency Rank, Market Basket и Item Comparison для владельцев Brand Registry. CX Codex +1 1 300установок

Из того же набора19

Весь набор →
product-marketing-ops Скилл product-marketing-ops для Claude Code помогает планировать запуски B2B SaaS и работу с позиционированием. Включается при подготовке GTM-стратегии, анализе конкурентов или выборе рынка для расширения. CX Codex +1 ★ 43звёзд репо partnerships-ops Скилл partnerships-ops для Claude Code помогает проектировать партнёрские программы и маркетинг через партнёров. Подключается к задачам о совместных кампаниях, интеграциях, маркетплейсах и канальных продажах. CX Codex +1 ★ 43звёзд репо paid-media-ops Скилл paid-media-ops для Claude Code помогает планировать и вести платную рекламу B2B SaaS. Разбирает кампании, конверсии, бюджеты и рекламные площадки. CX Codex +1 ★ 43звёзд репо marketing-project-mgmt Скилл marketing-project-mgmt для Claude Code координирует маркетинговые кампании, распределение ресурсов, спринты и работу с клиентами. Включается при планировании кампании, балансировке загрузки или подготовке QBR. CX Codex +1 ★ 43звёзд репо growth-ops Скилл growth-ops для Claude Code помогает проектировать рост через продукт и работу с текущими клиентами. Направляет запросы по активации, интерактивным демо, удержанию и расширению выручки профильным специалистам. CX Codex +1 ★ 43звёзд репо events-ops Скилл events-ops для Claude Code планирует B2B SaaS-мероприятия и вебинары. Включается при выборе конференций и спонсорств, разработке стратегии стенда или оценке event pipeline. CX Codex +1 ★ 43звёзд репо email-marketing-ops Скилл email-marketing-ops для Claude Code и Codex настраивает email-маркетинг, цепочки писем и доставляемость для B2B SaaS. Запускается при разработке рассылок, настройке DMARC или аудите частоты писем. CX Codex +1 ★ 43звёзд репо developer-marketing-ops Скилл developer-marketing-ops для Claude Code помогает прорабатывать DevRel и маркетинг B2B SaaS для разработчиков. Запускается на запросах о документации, SDK, quickstart или сообществе разработчиков. CX Codex +1 ★ 43звёзд репо

Спонсоры проекта

Компании, которые помогают держать аналитику открытой