Your achievements

Level 1

0% to

Level 2

Tip /
Sign in

Sign in to Community

to gain points, level up, and earn exciting badges like the new
Bedrock Mission!

Learn more

View all

Sign in to view all badges

Coffee Break: Join us on Wednesday, April 5th at 10am PT to chat with Ben Gaines, Director of Product Management for Adobe Analytics. Ben will be online to answer your Analytics and Adobe Summit questions.

Chatbot Depth Implementation


Level 1

I need to implement an analytics solution for a chatbot.


The chatbot is basically a predefined set of answers. Users only have the option of selecting their issue from the set of questions that we have provided


For eg:

User launches chatbot and is given following options:
1. issue with a refund

2. issue with an order

User selects an option and then he gets the next set of corresponding questions.


When he reaches the end of a specific flow we also give him an option of 'Do you want to ask another question ?'  wherein he gets to restart the entire process again. 


Currently I have a counter eVar (ie. eng_depth) implemented that increments each time an option in the chatbot is selected. 

the eng_depth resets to 1 each time a corresponding prop (that records the chatbot's response) is set to 'Do you want to ask another question ?' and another one that records the user's response is set to 'Yes'


As an analyst I need to be able to track, what extent a user goes deep into the chatbot, 

Here's a small example user chatbot journey:

Screenshot 2021-05-20 at 8.30.42 PM.png


Would also be great if you can guide on how I can have this as a metric so that i can get 'Average Chatbot depth' for a specific other Evar such as 'page_referrer' etc. that tells me on which page the user is most likely to have the most engaging chatbot session.


Can you please help ? 


1 Reply


Level 2
I think you can pass a Chat Bot ID which will be unique for each interaction. Based on chat bot Id you can breakdown all intent. And using calculated metrics you can identify the depth as well.