Back to all support articles
Analytics analyticsdashboardconversionoptimization
Analytics and optimization
Every chatbot has its own Dashboard. It answers two questions: are people using the bot, and does the bot help them buy?
Engagement metrics
| Metric | What it tells you |
|---|---|
| Chat Sessions | Unique conversations in the period |
| Messages | Total messages exchanged |
| Avg Messages | Messages per visitor — very low means people bounce, very high means answers are hard to reach |
| Session Duration | How long a conversation lasts |
| Engaged Session Rate | Share of sessions that went beyond a single message |
| Email Capture Rate | Share of sessions where a visitor left an email address |
Commerce metrics
| Metric | What it tells you |
|---|---|
| Chats with Products % | How often a conversation produced product suggestions |
| Impressions | Individual product cards shown |
| Ask AI clicks/taps | Visitors asking the bot for more detail about a specific product |
| Buy Now clicks/taps | Visitors going from the chat to a product page |
| Conversion Rate | Click-through rate after products were suggested |
| Purchase Chat % | Share of chat visitors that clicked Buy Now |
| Catalog Coverage | Share of your catalog the AI is actually able to suggest |
| Predefined (Canned) Response Rate | Share of replies served from your canned responses instead of the AI |
Breakdowns
- Language Distribution — check that your active languages match real demand before writing more content
- Device Types — desktop, mobile, tablet
- Top 10 most suggested Products and Top 10 Buy Now click/tap actions — the gap between these two lists is where your product data is letting you down
- Chat Activity Trends and Product Engagement Trends — 7-day and 30-day charts, with a Last 7 Days versus Previous 7 Days comparison
A weekly review that actually changes something
- Compare Last 7 Days with Previous 7 Days. Note anything that moved more than a few percent.
- Open Top 10 most suggested Products. For each product that gets many impressions but no Buy Now clicks, read the description in your feed. Thin or generic descriptions cause this.
- Open Chat conversations and filter on sessions with no product suggestions. Those are the questions your catalog cannot answer.
- Fix one thing: a product description, a missing category, a canned response, or a policy explanation in Site info → What does your website sell/do?
- Check the same numbers next week.
What to fix, in order
- Low Chats with Products % — check your search strategy and whether products imported for the visitor’s language. See Product recommendation settings.
- Low Conversion Rate — the right products appear but do not convince. Improve titles, descriptions, and images in your feed.
- Low Catalog Coverage — large parts of your catalog are invisible to the AI. Usually missing categories or missing
product_typevalues. - High Predefined Response Rate — many questions are being answered from canned text. Check that those canned answers still say the right thing.
- Low Engaged Session Rate — the opening message is not landing. Rewrite the welcome message and the attention popup in Chatbot and error messages.
Change one thing at a time
Every change to prompts, responses, or product data moves several metrics at once. Change one thing, wait a week, then read the same comparison again. That is the only way to know which change did the work.
Related articles
Review chat conversations
Search, filter, and annotate real conversations to find what your bot is getting wrong.
Account setup
Create your Mikabot account, verify your email, and find your way around the dashboard.
AI responses explained
The difference between canned and smart responses, and how the bot decides which to use.
Contact us
Our support team is ready to give you fast, personal help when you need it.