Three key data points should be tracked when considering the conversion effectiveness of chatbots. These data points are:
• Chatbot triggers:How many times website visitors encountered the chatbot. You can think of this just about the same way as impressions for ads. All impressions are not actual views, and the same goes for triggers.
• Bot conversations: How many times website visitors initiated a discussion with the chatbot. There’s no fluff here – a conversation is recorded after the user’s first interaction with the bot.
• Converted leads: How many people converted into a lead from the discussion. That is, people who submitted their information or completed another defined action. The criteria for a lead vary by chatbot use case. In many cases, a conversion is recorded when a user gives their name, phone number and/or email address, and company name.
By tracking these three metrics, you can get a good understanding of a chatbot’s performance. It’s a mini funnel, where triggers should lead to conversations and further on into conversions. Looking at chatbots this way makes it easy to analyze the performance and pinpoint any issues.
For example, a chatbot could have thousands of triggers every week, but only few conversations. This clearly indicates that the bot is not well-placed on the website, or that the opening line is not relevant to the users. Or both.
In another case, people may start chatting with a bot at a high rate, but they just don’t convert. This points to problems in conversation design and flow. The bot could be asking for conversion too bluntly, the conversation could be too long, or questions could be presented in the wrong order for users to stay motivated.