analytics_generate_action_items
Generate prioritized SEO and content action items by analyzing merged Google Search Console and Analytics data to identify optimization opportunities.
Instructions
Generate prioritized SEO and content action items from merged data.
Input Schema
TableJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| site_url | No | ||
| property_id | No | ||
| start_date | No | ||
| end_date | No | ||
| include_previous_period | No | ||
| max_rows | No | ||
| max_items | No | ||
| priorities | No |
Output Schema
TableJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
No arguments | |||
Implementation Reference
- src/seo_analytics_mcp/server.py:655-697 (handler)Main MCP tool handler function that generates prioritized SEO and content action items. It fetches merged page data from GSC and GA4, calls generate_action_items() to analyze pages, filters by optional priority levels, and returns a structured summary with counts and the prioritized action items list.
@mcp.tool() def analytics_generate_action_items( site_url: str | None = None, property_id: str | None = None, start_date: str | None = None, end_date: str | None = None, include_previous_period: bool = True, max_rows: int = 50000, max_items: int | None = None, priorities: list[str] | None = None, ) -> dict[str, Any]: """Generate prioritized SEO and content action items from merged data.""" settings = _get_settings() data = _fetch_page_data( site_url, property_id, start_date, end_date, include_previous_period=include_previous_period, max_rows=max_rows, ) items = generate_action_items(data["merged_pages"], settings, max_items=max_items) if priorities: allowed = {p.lower().strip() for p in priorities} items = [i for i in items if str(i.get("priority", "")).lower() in allowed] priority_counts = Counter(item["priority"] for item in items) category_counts = Counter(item["category"] for item in items) return { "ranges": data["ranges"], "site_url": data["site_url"], "property_id": data["property_id"], "summary": { "total_items": len(items), "priority_counts": dict(priority_counts), "category_counts": dict(category_counts), "portfolio": summarize_portfolio(data["merged_pages"]), }, "items": items, } - Core analysis function that iterates through merged page metrics, scores each page using score_page(), builds comprehensive action item objects with scores, priorities, categories, reasons, recommendations, and supporting evidence, then sorts by score/confidence and returns the top N items.
def generate_action_items( merged_pages: list[dict[str, Any]], settings: Settings, *, max_items: int | None = None, ) -> list[dict[str, Any]]: limit = max_items if max_items is not None else settings.default_max_action_items items: list[dict[str, Any]] = [] for page in merged_pages: result = score_page(page, settings) if result.score <= 0: continue item = { "url": page["url"], "score": result.score, "priority": result.priority, "category": result.categories[0] if result.categories else "opportunity", "categories": result.categories, "expected_impact": result.expected_impact, "effort": result.effort, "confidence": result.confidence, "reasons": result.reasons, "recommended_actions": result.recommendations, "evidence": { "gsc_impressions": round(float(page.get("gsc_impressions", 0.0)), 2), "gsc_clicks": round(float(page.get("gsc_clicks", 0.0)), 2), "gsc_ctr": round(float(page.get("gsc_ctr", 0.0)), 4), "gsc_position": round(float(page.get("gsc_position", 0.0)), 2), "ga4_sessions": round(float(page.get("ga4_sessions", 0.0)), 2), "ga4_engagement_rate": round( float(page.get("ga4_engagement_rate", 0.0)), 4 ), "ga4_conversion_rate": round( float(page.get("ga4_conversion_rate", 0.0)), 4 ), "gsc_clicks_delta_pct": page.get("gsc_clicks_delta_pct"), "ga4_sessions_delta_pct": page.get("ga4_sessions_delta_pct"), }, } items.append(item) items.sort(key=lambda i: (i["score"], i["confidence"]), reverse=True) return items[: max(1, limit)] - Page scoring engine that analyzes GSC (impressions, clicks, CTR, position) and GA4 (sessions, conversion rate, engagement) metrics along with period deltas. Identifies CTR optimization, conversion optimization, content refresh, and scale opportunities. Calculates composite scores, priority levels (high/medium/low), expected impact, effort estimates, and confidence levels.
def score_page(page: dict[str, Any], settings: Settings) -> ScoreResult: score = 0.0 categories: list[str] = [] reasons: list[str] = [] recommendations: list[str] = [] impressions = float(page.get("gsc_impressions", 0.0)) clicks = float(page.get("gsc_clicks", 0.0)) ctr = float(page.get("gsc_ctr", 0.0)) position = float(page.get("gsc_position", 0.0)) sessions = float(page.get("ga4_sessions", 0.0)) conversion_rate = float(page.get("ga4_conversion_rate", 0.0)) engagement_rate = float(page.get("ga4_engagement_rate", 0.0)) clicks_delta = page.get("gsc_clicks_delta_pct") sessions_delta = page.get("ga4_sessions_delta_pct") if impressions >= settings.min_impressions_for_ctr_action and ctr < settings.target_ctr: ctr_gap = (settings.target_ctr - ctr) / max(settings.target_ctr, 1e-6) ctr_score = min(45.0, ctr_gap * 45.0 + _log_scale(impressions, 6.0)) score += ctr_score categories.append("ctr_optimization") reasons.append( "High impressions with below-target CTR indicate snippet/title opportunity." ) recommendations.extend( [ "Rewrite title and meta description to better match dominant queries.", "Test stronger value proposition in the first 60 title characters.", ] ) if sessions >= settings.min_sessions_for_conversion_action and conversion_rate < settings.target_conversion_rate: cr_gap = (settings.target_conversion_rate - conversion_rate) / max( settings.target_conversion_rate, 1e-6, ) cr_score = min(45.0, cr_gap * 42.0 + _log_scale(sessions, 6.0)) score += cr_score categories.append("conversion_optimization") reasons.append( "Strong traffic with weak conversion efficiency suggests on-page UX/content friction." ) recommendations.extend( [ "Strengthen above-the-fold CTA and internal next-step links.", "Add trust proof and tighten informational-to-commercial transition sections.", ] ) if isinstance(clicks_delta, (int, float)) and clicks_delta <= -0.2: drop_score = min(30.0, abs(clicks_delta) * 60.0) score += drop_score categories.append("content_refresh") reasons.append("Organic clicks are declining versus the previous period.") recommendations.append( "Refresh outdated sections and compare SERP competitors for intent drift." ) if isinstance(sessions_delta, (int, float)) and sessions_delta <= -0.2: drop_score = min(30.0, abs(sessions_delta) * 55.0) score += drop_score categories.append("content_refresh") reasons.append("On-site sessions are declining versus the previous period.") recommendations.append( "Audit UX changes, page speed, and content relevance for recent traffic loss." ) if ( impressions >= settings.min_impressions_for_ctr_action and sessions >= settings.min_sessions_for_conversion_action and ctr >= settings.target_ctr and conversion_rate >= settings.target_conversion_rate ): scale_score = min(35.0, _log_scale(clicks + sessions, 7.0) + 10.0) score += scale_score categories.append("scale_winner") reasons.append("Page performs well in both acquisition and conversion.") recommendations.extend( [ "Expand topic cluster around this page's highest-performing query themes.", "Promote this page via internal links from adjacent intent pages.", ] ) if position > 8 and impressions > 0: score += min(12.0, (position - 8) * 1.5) reasons.append("Average position indicates page may be near page-one threshold.") if not categories and score <= 0: return ScoreResult( score=0.0, categories=[], reasons=[], recommendations=[], priority="low", expected_impact="low", effort="low", confidence=0.0, ) unique_categories = sorted(set(categories)) unique_reasons = list(dict.fromkeys(reasons)) unique_recommendations = list(dict.fromkeys(recommendations)) sources = 0 if impressions > 0: sources += 1 if sessions > 0: sources += 1 volume_factor = 0.0 if impressions > 0: volume_factor += min(0.25, math.log10(impressions + 1) / 10) if sessions > 0: volume_factor += min(0.25, math.log10(sessions + 1) / 10) confidence = min(1.0, 0.35 + 0.2 * sources + volume_factor) return ScoreResult( score=round(score, 2), categories=unique_categories, reasons=unique_reasons, recommendations=unique_recommendations, priority=_priority_from_score(score), expected_impact=_expected_impact(score), effort=_effort(unique_categories), confidence=round(confidence, 2), ) - src/seo_analytics_mcp/server.py:270-270 (registration)Tool name registered in the capabilities() function list of available tools, which exposes it to MCP clients as a callable tool.
"analytics_generate_action_items", - Settings dataclass defining configuration parameters used by the action item generator, including min_impressions_for_ctr_action, min_sessions_for_conversion_action, target_ctr, target_conversion_rate, and default_max_action_items thresholds that control when action items are generated.
@dataclass(frozen=True) class Settings: enable_gsc: bool enable_ga4: bool require_explicit_gsc_site_url: bool default_gsc_site_url: str | None default_ga4_property_id: str | None default_lookback_days: int canonical_base_url: str | None min_impressions_for_ctr_action: int min_sessions_for_conversion_action: int target_ctr: float target_conversion_rate: float default_max_action_items: int