Analytics

Metrics Glossary

Understand your chatbot's performance metrics and analytics.

What are Metrics?

The Metrics dashboard gives you an overview of your chatbot's activity and performance over a selected period.

Use it to monitor conversations, response times, technical errors, handoffs, and customer sentiment.

Metrics

Conversation Metrics

These metrics show how frequently customers interact with your chatbot.

MetricWhat it means
Total ConversationsThe number of conversations recorded during the selected period.
Messages SentThe number of messages sent by your chatbot.
Messages ReceivedThe number of messages received from users.
Total MessagesThe combined number of messages sent and received.

Response and Reliability

Use these metrics to understand how quickly and reliably your chatbot operates.

MetricWhat it meansWhat to look for
Error RateThe proportion of activity associated with technical errors.An increase may indicate failed requests or other technical issues.
Average Response TimeThe average time your chatbot takes to respond.Longer response times may indicate slower processing or increased traffic.

Note: Technical errors are different from incorrect answers. A chatbot can respond successfully but still provide inaccurate information.

Handoff Metrics

MetricWhat it meansWhat to look for
Handoff RateHanded-off conversations as a percentage of all conversations in the selected period. Handoffs that do not point to a gap in your bot's knowledge, such as those made outside business hours, are not counted.A high rate may indicate that customers frequently need human assistance. Review handed-off conversations to identify recurring questions or gaps in your bot's knowledge.

A handoff is not necessarily a bad outcome. Some inquiries are better handled by a human agent.

Sentiment Metrics

Sentiment metrics help you understand the tone of customer interactions based on the sentiment information available from conversations.

MetricWhat it meansWhat to look for
Average SentimentThe average sentiment score across the available conversation data for the selected period.Look for changes in sentiment over time and investigate conversations that may explain them.
Sentiment AnalysisA more detailed view of sentiment across conversations.Explore sentiment patterns and review individual conversations for context.

Sentiment scores are estimates and may not reflect every customer's full experience. Review the actual conversations before drawing conclusions.

Activity Metrics

MetricWhat it meansWhat to look for
Activity TrendA visualization of chatbot or user activity over time.Identify busy periods, quieter periods, and changes in traffic. Compare these patterns with response times, errors, and handoffs.

Using Metrics to Improve Your Chatbot

Use your metrics to identify areas that may need attention.

If you notice...Consider...
High conversation volumeChecking whether your bot handles common questions effectively.
An increasing error rateContacting support if the errors continue.
Longer response timesChecking whether traffic or processing time has increased.
Frequent handoffsReviewing conversations and improving relevant training data.
Changes in sentimentReading conversations to understand the customer experience.

Metrics help you identify patterns, but individual conversations provide important context for understanding why those patterns occur.

You may also want to check out these guides

Did this fix it?

If you are unable to solve your issue, tell us what happened. You do not need to be signed in.

No, message us