7 Chatbot Metrics Every E-shop Owner Should Track

You've deployed an AI chatbot on your e-shop. Great. But how do you know if it's actually working? Vanity metrics like "total conversations" don't tell the full story. Here are the 7 metrics that actually matter.

1. Resolution Rate

What it measures: The percentage of conversations where the chatbot fully resolved the customer's question without human intervention.

This is the single most important metric for your chatbot. A well-trained AI chatbot should resolve 85–98% of conversations on its own. If your resolution rate is below 80%, it usually means the chatbot needs more training data or better instructions.

Target: 90%+ for a mature deployment.

2. Customer Satisfaction Score (CSAT)

What it measures: How satisfied customers are with the chatbot interaction, typically measured via a thumbs up/down or 1–5 star rating at the end of the conversation.

High resolution rate means nothing if customers hate the experience. CSAT tells you whether the chatbot is actually helpful or just technically answering questions in a way that frustrates people.

Target: 85%+ positive ratings.

3. Average Response Time

What it measures: How quickly the chatbot responds to each customer message.

One of the biggest advantages of AI is speed. Your chatbot should respond in under 3 seconds. If response times are creeping up, it could indicate performance issues or overly complex processing chains.

Target: Under 2 seconds for most queries.

4. Escalation Rate

What it measures: The percentage of conversations that get handed off to a human agent.

Some escalation is healthy — you want complex issues handled by humans. But a high escalation rate (above 20%) suggests the chatbot isn't trained well enough or is encountering questions it can't handle.

Track why conversations escalate. Common reasons include:

• Missing product information
• Complex return/refund scenarios
• Customer frustration (they asked for a human)
• Questions outside the chatbot's scope

Target: Under 15% for a well-trained chatbot.

5. Conversion Influence

What it measures: How many customers who interacted with the chatbot went on to make a purchase.

This is where the chatbot proves its ROI. Compare the conversion rate of visitors who used the chatbot versus those who didn't. In most e-shops, chatbot users convert 2–4x higher than non-chatbot visitors.

Why? Because customers who engage with the chatbot are getting their questions answered, their objections handled, and their confidence boosted — all of which lead to purchases.

Target: Chatbot users should convert at least 2x higher than average site visitors.

6. Fallback Rate

What it measures: How often the chatbot responds with "I don't know" or a generic fallback message.

Every fallback is a missed opportunity. If the chatbot can't answer a question, the customer is left without help — and likely leaves your store. Track which questions trigger fallbacks and use this data to continuously improve your chatbot's knowledge base.

Target: Under 5% fallback rate.

7. Cost Per Resolution

What it measures: How much each resolved conversation costs you — combining chatbot subscription, human agent time for escalated queries, and any associated tools.

This is the metric your CFO cares about. Calculate it like this:

A typical AI chatbot reduces cost per resolution from €5–€15 (human only) to €0.10–€0.50 (AI + human hybrid). That's a 90%+ reduction.

Target: Under €1 per resolution for the blended rate.

How to Use These Metrics

Don't just track these numbers — act on them. Set up a weekly review:

1. Check resolution and fallback rates. If they're trending wrong, review recent conversations to find gaps.
2. Read escalated conversations. Each one is a learning opportunity — can the chatbot handle this next time?
3. Monitor CSAT trends. A sudden drop usually signals a specific problem you can fix quickly.
4. Compare conversion influence monthly. This is your proof that the chatbot is driving revenue, not just saving costs.