📊 Metrics & Self-Improvement

BaC Principle

“A process that doesn’t measure itself can’t improve itself. Every KPI is a feedback signal. Every signal drives a proposal. Every proposal closes the loop.” — Business as Code Manifesto

Stage: S9 — Improvement Loop Type: AUTO (data collection + proposals) + HUMAN (proposal approval) Cadence: Weekly data collection; weekly metrics review; monthly strategy review

See Map of Content for full vault navigation.

🔄 Improvement Loop Architecture

flowchart TD
    DATA["📥 Data Collection<br/>GSC + Ahrefs + AI citation probes"]
    AGG["📊 Metric Aggregation<br/>SEO track + AEO track"]
    ANOM["🚨 Anomaly Detection<br/>Automated rules"]
    PROP["📝 Improvement Proposals<br/>Generated by MetricsAgent"]
    REVIEW["👤 Human Review<br/>~30 min/week"]
    DEPLOY["🚀 Deploy Changes<br/>Update manifest / agents / target spec"]

    DATA --> AGG --> ANOM --> PROP --> REVIEW --> DEPLOY --> DATA

📈 SEO Track KPIs

Primary Dashboard

Metric	Source	Target	Alert Threshold
Organic sessions/week	GSC	Growing 10%+ MoM	Drop > 15% WoW
Avg SERP position (all keywords)	GSC	Improving	Regression > 3 positions WoW
Avg SERP position (target clusters)	GSC	≤ 15	> 25 → deprioritized cluster
Organic CTR	GSC	≥ 3%	< 2% → title/meta review
Indexed pages	GSC	Growing	Stagnation > 2 weeks
Referring domains	Ahrefs	Growing	Flat for 30 days
New backlinks/month	Ahrefs	≥ 5	0 → authority building alert
Core Web Vitals pass rate	PageSpeed	≥ 75%	< 60% → dev escalation
Pages with content score ≥ 75	SurferSEO	≥ 80% of indexed	< 65% → content optimization sprint

Funnel Metrics

Stage	Metric	Week N	Week N-1	Delta
Discovery	Impressions	—	—	—
Click	Organic clicks	—	—	—
Engagement	Avg time on page	—	—	—
Conversion	Goal completions	—	—	—

🤖 AEO Track KPIs

AI Citation Dashboard

Metric	Source	Target	Alert Threshold
AI citation rate (probed queries)	LLM probe	≥ 20%	< 10% → AEO sprint
Google AI Overview appearances	GSC (AI)	≥ 5 queries	Declining → schema review
Perplexity citation count/month	Manual probe	Growing	Drop > 20% MoM
ChatGPT citation rate	OpenAI API probe	≥ 15%	< 5% → content gap issue
FAQ schema rich result rate	GSC rich results	≥ 60% of eligible pages	< 40% → schema fix
Avg AEO readiness score (site-wide)	AEO agent output	≥ 75/100	< 60 → AEO agent re-run
Citation gap queries addressed	Content inventory	100% of top 10 gaps	Any unaddressed → priority

Weekly AI Citation Probe Protocol

citation_probe:
  frequency: weekly
  queries_per_run: 30
  platforms:
    - chatgpt: 10 queries (OpenAI API, model: gpt-4o)
    - perplexity: 10 queries (Perplexity API)
    - claude: 5 queries (Claude API, model: claude-opus-4-6)
    - gemini: 5 queries (Gemini API)
 
  query_categories:
    - definitional: "What is [target concept]?"
    - instructional: "How to [target action] with AI?"
    - comparative: "What is the best tool for [task]?"
    - brand: "What is Business as Code / businessascode.co?"
 
  output:
    - cited: true/false
    - citation_url: if cited
    - competitor_cited_instead: domain
    - answer_gap: what the model answered that we don't cover

🚨 Anomaly Detection Rules

anomaly_rules:
  - id: ANO-01
    name: traffic_drop
    condition: "organic_sessions this week < 85% of 4-week rolling average"
    action: Trigger immediate audit — check GSC for manual actions, algorithm updates, indexing drops
    notify: CMO
 
  - id: ANO-02
    name: ranking_regression
    condition: "avg SERP position for target clusters degrades > 5 positions WoW"
    action: Pull affected keywords; compare to competitor SERP changes; check for content updates needed
    notify: CMO
 
  - id: ANO-03
    name: indexing_stall
    condition: "indexed pages count unchanged for 14 days"
    action: Check sitemap submission status; check robots.txt; re-submit sitemap via GSC
    notify: CMO
 
  - id: ANO-04
    name: aeo_citation_drop
    condition: "AI citation rate drops > 25% MoM"
    action: Re-run AEO structuring agent on top 10 pages; check for schema validation errors
    notify: CMO
 
  - id: ANO-05
    name: gate_rejection_spike
    condition: "Gate #2 rejection rate > 30% of sampled pages for 2 consecutive weeks"
    action: Review content optimization engine prompt; likely quality regression
    proposal: Retune content brief template or Claude API system prompt
 
  - id: ANO-06
    name: backlink_stall
    condition: "zero new referring domains for 30 days"
    action: Review authority building agent output; check if outreach queue is being executed
    notify: CMO

📝 Improvement Proposal Format

Every proposal generated by the MetricsAgent follows this schema:

improvement_proposal:
  proposal_id: "IMP-2026-04-001"
  generated_at: "2026-04-09"
  generated_by: MetricsAgent
  triggered_by: ANO-04   # or metric name, or human feedback from gate
 
  problem: >
    AI citation rate dropped from 18% to 9% over the past 3 weeks.
    Probe analysis shows Perplexity is now citing zapier.com for
    "business automation AI" queries that we previously captured.
 
  proposed_change:
    target: AEOStructuringAgent
    change_type: prompt_update
    description: >
      Add explicit entity relationship mapping to AEO pass:
      "Business as Code" → related to → [AI agents, process automation, BPM].
      This signals to LLMs how our content fits their knowledge graph.
 
  test_plan:
    method: a_b_test
    variant_a: current AEO pass (control)
    variant_b: AEO pass with entity relationship section
    metric: ai_citation_rate
    minimum_sample: 4 weeks
    success_threshold: "Citation rate returns to ≥ 18%"
 
  human_decision_required: true
  urgency: high
  estimated_effort: low

🧪 A/B Test Registry

ab_tests:
  - test_id: AB-001
    element: content_title_format
    variant_a: "What is [Topic]? Complete Guide (Year)"
    variant_b: "[Topic]: Definition, Examples & How It Works"
    metric: organic_ctr
    started: 2026-04-02
    status: running
    minimum_pages: 10
    results: pending
 
  - test_id: AB-002
    element: faq_question_count
    variant_a: "5 FAQ questions per page"
    variant_b: "10 FAQ questions per page"
    metric: aeo_citation_rate
    started: 2026-04-02
    status: running
    minimum_pages: 8
    results: pending

A/B Promotion Rules:

Minimum 4 weeks per test
Statistical significance: ≥ 90% confidence
Minimum sample: ≥ 8 pages per variant
Winning variant becomes default in the relevant agent spec

📅 Review Cadences

review_cadences:
  weekly:
    duration: 30 min
    owner: CMO
    agenda:
      - Review SEO + AEO dashboard (5 min)
      - Check anomaly alerts (5 min)
      - Review improvement proposals from MetricsAgent (10 min)
      - Approve / reject / defer proposals (10 min)
 
  monthly:
    duration: 60 min
    owner: CMO
    agenda:
      - Full funnel review (10 min)
      - A/B test results review (10 min)
      - Target Definition update — add/remove keyword clusters (15 min)
      - Content calendar planning for next month (15 min)
      - Agent prompt review — any systemic quality issues (10 min)

🗒️ Improvement Log

improvement_log:
  - version: 1.0.0
    date: 2026-04-02
    change: Initial process definition
    impact: Baseline established

📊 Agent Reporting Matrix

Agent	Reports	Frequency
SiteAuditAgent	Audit report + baseline metrics	Monthly
KeywordResearchAgent	Keyword universe + AEO gap analysis	Monthly
ContentOptimizationEngine	Content quality scores + batch completion	Weekly
AEOStructuringAgent	AEO readiness scores + schema validation	Weekly
TechnicalSEOAgent	Issues resolved + CWV status	Weekly
AuthorityBuildingAgent	Citation rate + backlink acquisition	Weekly
MetricsAgent	KPI dashboard + anomaly alerts + proposals	Weekly

01-Process-Manifest — S9 stage definition + change protocol
02-Target-Definition — Primary optimization targets; updated by improvement proposals
06-AEO-Structuring-Agent — AEO readiness scores; agent prompts updated via proposals
05-Content-Optimization-Engine — Content quality scores; briefs updated via proposals
09-Human-Review-Gates — Gate feedback feeds anomaly detection
08-Authority-Building-Agent — Citation rates tracked here

Explorer

Metrics & Self-Improvement

📊 Metrics & Self-Improvement

🔄 Improvement Loop Architecture

📈 SEO Track KPIs

Primary Dashboard

Funnel Metrics

🤖 AEO Track KPIs

AI Citation Dashboard

Weekly AI Citation Probe Protocol

🚨 Anomaly Detection Rules

📝 Improvement Proposal Format

🧪 A/B Test Registry

📅 Review Cadences

🗒️ Improvement Log

📊 Agent Reporting Matrix

Graph View

Table of Contents

Backlinks

Explorer

Metrics & Self-Improvement

📊 Metrics & Self-Improvement

🔄 Improvement Loop Architecture

📈 SEO Track KPIs

Primary Dashboard

Funnel Metrics

🤖 AEO Track KPIs

AI Citation Dashboard

Weekly AI Citation Probe Protocol

🚨 Anomaly Detection Rules

📝 Improvement Proposal Format

🧪 A/B Test Registry

📅 Review Cadences

🗒️ Improvement Log

📊 Agent Reporting Matrix

📎 Related Files

Graph View

Table of Contents

Backlinks