Service Analytics Guide: Monitoring and Analyzing Customer Service Performance
Summary
One-line takeaway: Service Analytics 2.0 upgrades monitoring with deeper service, conversion, and risk insights.
- Core Value/Change: ① Redesigned dashboard with live and historical views ② richer bot + agent metrics ③ stronger drill-down and export tools
- Actionable Steps: ① Open Customer Messages ② click Service Analytics icon ③ use Overview, Operation, Agent, All Chats, and Insights pages
- Index & Logic: ① Conversion tracks orders within 7 days after chat ② violation rate = risky chats / total chats ③ response and satisfaction formulas are defined
- Immediate Next Action: Open Service Analytics and review Overview metrics
- Comprehensive overview dashboard — See your shop's real-time and historical service performance at a glance
- Expanded metrics — Track chatbot and human agent data side by side, including chat volume breakdowns, satisfaction drivers, and sales conversion
- Drill-down analysis — Break down key metrics by intent category (Pre-purchase, Aftersales, Logistics, Other) and by individual agent
- Risk monitoring — Identify risky or policy-violating conversations and the agents involved
- Exportable reports — Download charts and tables for offline analysis and record-keeping
How to Access Service Analytics
- Go to Customer Messages
- Click the Service Analytics icon in the top-right corner
- You will land on the new Operation Overview dashboard.
- You will land on the new Operation Overview dashboard.
Operation
Overview
The Overview is the first screen you see when you open Service Analytics. It shows high-level performance metrics for your shop, split into two areas:
Live overview (refreshes every minute):- Sessions today
- Online agents
- Agents with active chats
- Sessions volume — Total sessions and breakdown by type (AI-only, Human-only, AI-to-Human)
- Service performance — Customer Satisfaction, 24h Response Rate, Average Response Time
- Conversion — Revenue After Inquiry, Orders After Inquiry, Conversion Rate
- Violation — Violating Conversations, Affected Agents, Violation Rate
Operation section contains three sub-pages that provide detailed analysis of your shop's chat operations, service quality, and sales impact.
Chat Volume
This page shows a complete breakdown of your shop's inbound chat activity.
Data overview cards:
- Total chats received in the selected period
- AI Bot-only chats — Conversations handled entirely by the automated assistant
- Human-only chats — Conversations handled entirely by a human agent
- AI Bot-to-human transfer chats — Conversations that started with the bot and were transferred to a human agent
Drill-down by intent:
A table breaks down chat volume by first-level intent category: Pre-purchase, Aftersales, Logistics, and Other. This table can be downloaded for offline analysis.
Daily trend chart:
A line chart shows the daily trend for total chats and each chat type, so you can spot patterns and volume spikes.
Service Performance
This page helps you understand the quality and timeliness of your shop's customer responses.
Key metric cards:
- Customer Satisfaction
- Drill down by negative review reason
- Click to view negative-rated chat details (redirects to All Chats with the "Negative rating" filter applied)
- 24-hour response rate
- An educational banner appears for first-time viewers, explaining how this metric impacts conversion and retention
- Industry benchmark data is shown in grey text (e.g., "Top X% of shops achieve Y%")
- Period-over-period change is shown in green (improvement) or red (decline)
- Average agent response time
- Calculated as: first response time + average multi-turn response time
Hover over the ? icon next to any metric to see its full definition.
Drill-down by intent:
A bar chart breaks down each metric by intent category (Pre-purchase, Aftersales, Logistics, Others). Click any bar to jump to the All Chats page with the corresponding filter applied.
Daily trend charts:
Separate trend lines for 24-hour response rate, satisfaction rate, and average response time, plotted by day.
All charts support export/download.
Conversion
This page shows how your customer chat interactions translate into revenue.
Key metric cards:
- Revenue after inquiry — Total payment amount from post-inquiry orders
- Orders after inquiry — Number of orders placed after a customer chat
- Inquiry conversion rate — Customers who ordered / Total customers who chatted
- Period-over-period change displayed in green or red
Hover over the ? icon next to any metric to see its full definition.
Daily trend chart:
A line chart tracks conversion rate over time, plotted by day.
Agent
Agent Data contains two sub-pages for managing individual agent performance and team composition.
Team Performance
Real-time indicators (Today):
- Total agents count
- Online agents count
- Agents with active chats count
Agent Detail table:
A detailed table showing each agent's metrics for the selected time period. The Team Average row is pinned at the top of the table in bold for easy comparison.
Columns include:
Column | Description |
Agent name | Name and current online status |
| Status | Status of the agent |
Satisfaction rate | Percentage of positive ratings |
Avg. response time | Average time to respond |
24h response rate | Percentage of chats responded to within 24 hours |
Handled Chat | Total conversations handled |
Pending replies | Chats awaiting agent response |
You can filter by agent name, sort by any column, and download the full table.
Team Management
This page consolidates your team management functions (previously located in a separate module). You can manage agent roles, permissions, and assignments from this centralized location.
Chat History
All Chats
The All Chats page lets you browse, filter, and export individual conversation records.
Enhanced filters:
In addition to existing filters (date range, agent name, customer name, chat type, intent), the following new filters are available:
- Chat handling type — Filter by Bot-only or Human (including bot-to-human transfers)
- Rating level — Filter by star rating (1-5 stars). Note: In the US market, 3 stars is a neutral rating; in other markets, 3 stars is considered negative.
- Negative rating reason — Filter by the specific reason a customer gave a low rating
- Risky chat — Filter to show only conversations flagged for potential policy violations
Chat details table:
Default columns: Chat ID, Agent name, Customer name, Intent, Chat handling type, CSAT rating, Risky chat flag.
Click the Export button in the top-right corner to download the full table as a report.
Insights
Service Insights is a new module dedicated to proactive monitoring and risk management.
Risks
This page helps you identify and manage policy-violating or risky conversations.
Key metric cards:
- Risky Conversations — Total number of flagged conversations
- Affected Agents — Number of agents associated with risky chats
- Violation rate — Risky chats / Total chats
Drill-down charts:
- Bar chart: Risky chats broken down by intent category
- Bar chart: Risky chats broken down by agent
- Trend chart: Daily trend of risky chat volume
Flagged agent details:
A table listing each agent associated with risky conversations, with columns for agent name and total number of flagged chats. You can filter by agent name.
Note: Risky messages are automatically blocked from being sent to customers. When a message is blocked, the agent receives a system notification explaining the violation.
Negative Reviews
This page (previously a standalone section) is now housed under Insights. It provides a dedicated view of conversations that received negative customer ratings, helping you identify patterns and areas for service improvement.
Key Metrics Reference
Metric | Definition |
Total chats | All customer conversations received in the selected period |
Bot-only chats | Chats fully resolved by the automated assistant |
Human-only chats | Chats handled entirely by a human agent |
Bot-to-human transfer | Chats started by bot and escalated to a human agent |
24-hour response rate | Chats responded to within 24 hours / Total chats requiring response |
Satisfaction rate | Chats rated 4-5 stars / Total rated chats |
Average response time | First response time + average multi-turn response time |
Inquiry conversion rate | Customers who purchased within 7 days after chatting / Total chatting customers |
Orders after inquiry | Orders placed within 7 days of a customer chat |
Revenue after inquiry | Payment amount from post-inquiry orders |
Violation rate | Risky chats / Total chats |
FAQs
Question | Answer |
Where can I access Service Analytics? | Go to Customer Messages in TikTok Shop Seller Center and click the Service Analytics icon in the top-right corner. |
What time periods can I analyze? | You can select custom date ranges using the date picker at the top of each page. The selected dates are displayed for clarity. |
Can I export data from Service Analytics? | Yes. Most tables and charts include a download/export button. Click it to save the data for offline use. |
What is the difference between bot-only chats and bot-to-human transfer chats? | Bot-only chats are conversations fully resolved by the automated assistant without human involvement. Bot-to-human transfer chats started with the bot but were escalated to a human agent. |
Why do I see a 12-hour response rate instead of 24-hour? | Shops in Southeast Asian (SEA) markets use a 12-hour response rate standard instead of 24 hours. |
What counts as a "risky" conversation? | A risky conversation is one flagged by the system for containing content that may violate platform communication policies. Risky messages are automatically blocked before reaching the customer. |
How is the satisfaction rate calculated? | It is the percentage of customer-rated chats that received a 4 or 5 star rating out of all rated chats. |
How is the inquiry conversion rate calculated? | It is the number of customers who placed an order within 7 days after a chat, divided by the total number of customers who initiated a chat. |
What does the "Team Average" row in Agent Performance mean? | It shows the average value for each metric across all agents on your team, giving you a benchmark for individual performance comparison. |
I don't see the new Service Analytics 2.0 features yet. When will I have access? | Service Analytics 2.0 is being rolled out progressively. If you don't see the updated interface yet, check back soon — your shop will receive the update as the rollout continues. |
Can I see which specific agents are involved in risky conversations? | Yes. The Risk Conversations page includes an agent breakdown chart and a flagged agent details table that lists each agent and their number of risky chats. |
How can I improve my 24-hour response rate? | Ensure agents are online during peak hours, use automated replies to handle common questions, and monitor the Agent Performance table to identify agents who may need additional support or training. |