# Smart Summary
A feature that automatically analyzes heatmap data with AI and presents key page insights. Even without specialized heatmap knowledge, you can quickly understand your page's situation.
***
## About This Feature
Traditional heatmap analysis required switching between multiple heatmap types, checking metrics for each block individually, and deriving insights manually.
Smart Summary automates this analysis process with AI. It analyzes page data from 3 perspectives and presents major findings and action directions.
Using Smart Summary, you can discover:
* A summary of attention and exit points for users on the page
* Performance comparisons by advertising channel
* Behavioral trends across visitor segments
***
## How to Use
Click the **"Smart Summary"** tab in the heatmap toolbar to display AI analysis results in the right panel.
The right panel contains 3 sub-tabs:
Smart Summary screen (right panel with key insights and 3 sub-tabs)
***
## 3 Analysis Modules
### User Behavior Analysis
Analyzes user behavior on the page in block-level detail.
#### What You'll Learn
* **Key Insights**: Overall analysis conclusion (e.g., "36% exit at first view, suggesting insufficient specificity in value proposition")
* **Phase-by-Phase Behavior Analysis**: Breaks the page into 4 psychological stages and explains user behavior at each
* **Attention Points**: High-interest areas and their reasons
* **Exit Points**: Common exit locations and potential causes
#### Automatic Page Type Identification
When conducting user behavior analysis, Smart Heatmap first **automatically identifies the page type**. Since visitor intent and browsing patterns differ by page type, the analysis framework optimizes to match the page type.
| Page Type | Characteristics |
| ------------------------- | ------------------------------------------------------------- |
| Ad Landing Page | No navigation, single CTA, long-form text layout |
| Article Landing Page | Editorial article-style layout, educational content, soft CTA |
| Product Detail Page (PDP) | Buy box, user reviews, product gallery |
| Homepage | Full navigation, hero banner, brand story |
| Campaign Page | Event-driven, coupons/countdowns, time sales |
| Other | Login pages, contact forms, etc. |
#### 4-Stage Psychological Model
Based on the identified page type, block content is classified into **4 psychological stages** for analysis. The basic framework for the 4 stages is:
| Stage | User's Question | Meaning |
| ------- | --------------------------------- | ---------------------------------------------- |
| Phase 1 | "Is this relevant to me?" | Establishing relevance and first impression |
| Phase 2 | "Do you understand my challenge?" | Building empathy, reducing comprehension cost |
| Phase 3 | "Why should I buy this?" | Presenting rationale and proof, building trust |
| Phase 4 | "Is it really safe?" | Reducing risk, encouraging action |
#### Phase Names by Page Type
With the same Phase 1, a PDP user checks "Do I have the information needed to buy?" while an article LP user decides "Is it worth continuing to read?" Therefore, **phase names are optimized by page type**.
| Page Type | Phase 1 | Phase 2 | Phase 3 | Phase 4 |
| -------------------- | -------------------------- | --------------------------- | ------------------------------- | ------------------------------- |
| Ad Landing Page | Establish First Impression | Build Empathy | Give Purchase Reason | Ease Concerns |
| Article Landing Page | Build Reading Trust | Strengthen Problem Empathy | Convince with Solutions | Lower Action Barriers |
| Product Detail Page | Gather Purchase Info | Convey Product Value | Show Trust Basis | Push Purchase Action |
| Homepage | Clarify Brand and Target | Guide to Next Steps | Strengthen Trust and Rationale | Connect to Ongoing Relationship |
| Campaign Page | Focus on Campaign Value | Explain Participation Rules | Enable Quick Category Selection | Push to Order Completion |
> Phases 1-4 are universal alignment keys across all page types. Phase names are display labels matched to page type; analysis reports use names corresponding to the page type.
This classification is not simply dividing the page from top to bottom, but automatically determined by **the psychological role each block's content plays**. Since specialized classification logic is applied per page type, the same "reviews" block becomes Phase 3 (showing trust basis) for PDPs but Phase 3 (giving purchase reason) for ad LPs, reflecting contextual relevance.
#### Main Metrics Used
| Metric | Role |
| ----------------------- | ------------------------------------------------ |
| Block Average Stay Time | Determining user interest level |
| Block Exit Rate | Identifying exit points |
| Impression Rate | Confirming block reach status |
| Block Conversion Rate | Supporting judgment on content's CV contribution |
***
### Ad Effectiveness Analysis
Compares and analyzes performance of advertising channels driving traffic to the page.
> **About the conversion-rate basis**: Conversion rates in this analysis are calculated using **only the Primary Conversion** — the conversion at the top of the conversion settings panel. Even with multiple conversion goals configured, only the Primary Conversion is referenced here. To change which conversion drives the analysis, drag the target conversion to the top in the conversion settings panel. See [Block & Element Setup — Conversion Settings](/en/insight/smart-heatmap/block-element-settings.md#conversion-settings).
#### What You'll Learn
* **Key Insights**: Comparison of highest-performing and lowest-performing ad channels (e.g., "Google search achieves 2.1% conversion rate, most effective; Instagram ads at 0.3% need improvement")
* **Overall Overview**: Traffic composition and main ad channel performance summary
#### Analysis Mechanism
Ad effectiveness analysis evaluates each advertising channel on **2 axes: session count and conversion rate**.
| Session Count | Conversion Rate | Evaluation | Recommended Action |
| ------------- | --------------- | ---------------------- | ----------------------------------------------------------------- |
| High | High | Effective Channel | Maintain and strengthen; expand success factors to other channels |
| Low | High | High-Potential Channel | Carefully increase distribution volume |
| High | Low | Requires Attention | Verify landing page alignment; reduce if not improved |
| Low | Low | Under Investigation | Continue small-scale testing; assess effectiveness |
> When a specific channel represents over 95% of traffic, it's treated as overall overview rather than comparative analysis.
#### Main Metrics Used
| Metric | Role |
| ---------------------- | -------------------------- |
| Session Count (visits) | Traffic volume |
| Conversion Rate | Channel quality evaluation |
| Bounce Rate | Landing page alignment |
| Click Rate | Page engagement |
| CTA Click Rate | Action intent indicator |
| Average Stay Time | Content involvement |
***
### Audience Analysis
Analyzes behavioral trends across visitor segments (region, device, visitor type, etc.).
> **About the conversion-rate basis**: Like Ad Effectiveness Analysis, conversion rates here are based on **only the Primary Conversion** (the conversion at the top of the conversion settings panel).
#### What You'll Learn
* **Key Insights**: Comparison of highest-quality and improvement-needed visitor segments (e.g., "Repeat visitors convert at 3x the rate of new visitors; consider strengthening repeat visitor initiatives")
* **Overall Overview**: Visitor composition and main segment performance summary
#### Analysis Mechanism
Audience analysis uses the same **session count × conversion rate 2-axis approach** as ad effectiveness analysis to evaluate each segment.
Example analyzable segments:
| Segment Type | Examples |
| ----------------- | ----------------------------------------- |
| Visitor Type | New visitors, repeat visitors |
| Device | Mobile, PC, Tablet |
| Region | Country, prefecture |
| Traffic Source | Organic search, campaigns, direct traffic |
| Bounce/Non-Bounce | Bounce visitors, non-bounce visitors |
#### Main Metrics Used
Uses the same metric set as ad effectiveness analysis (excluding CTA click rate).
***
## Analysis Tips
### Use Smart Summary as Your Starting Point
Smart Summary is perfect as an **analysis starting point**. Based on AI-generated insights, dig deeper into areas of interest using other heatmap tabs.
| Smart Summary Discovery | Next Steps |
| ---------------------------------------- | -------------------------------------------------------------------------------------------------- |
| "High exit in Phase 1" | → Check [Exit Heatmap](/en/insight/smart-heatmap/exit-heatmap.md) around first view in detail |
| "Specific block shows high attention" | → Check detailed [Stay Heatmap](/en/insight/smart-heatmap/stay-heatmap.md) for that block |
| "Specific ad channel has low conversion" | → Use [Toolbar segmentation](/en/insight/smart-heatmap/toolbar.md) to check that channel's heatmap |
| "Repeat visitors have high conversion" | → Create initiatives for repeat visitors in Experience |
### Important Notes
* AI analysis results show "possibility hints," not definitive causal relationships
* Analysis for segments with low visit counts should be treated as reference only
* Regular review reveals time-series changes
***
## Related Features
* [Heatmap Comprehensive Guide](https://gitlab.com/PtmindDev/ptx/ptengine-helps/-/blob/main/DOCS/en/insight/smart-heatmap/overview.md) — Overall view of 6 features
* [Stay Heatmap](/en/insight/smart-heatmap/stay-heatmap.md) — For detailed attention point review
* [Exit Heatmap](/en/insight/smart-heatmap/exit-heatmap.md) — For detailed exit point review
* [Toolbar Features](/en/insight/smart-heatmap/toolbar.md) — For narrowed-down segment analysis
* [Metrics Reference](/en/insight/smart-heatmap/metric-reference.md) — Detailed metric definitions and formulas
---
# Agent Instructions: Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.
Perform an HTTP GET request on the current page URL with the `ask` query parameter:
```
GET https://helps.ptengine.com/en/insight/smart-heatmap/smart-summary.md?ask=
```
The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.
Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.