If you’re a bot developer and have questions about how to increase user engagement, stay tuned and keep reading because user engagement is the one metric that determines your bot’s destiny.
PennyCat is one of the members of Botanalytics. It allows people to find coupons for online stores. When a user asks to find coupon codes for a specific store, PennyCat makes a search on hundreds of coupon websites and provides all the available coupon codes found for that store. Other than this, it allows you to play a questions/answer game.
So, diving deep into the conversations is highly critical to explore the engagement of the different contents that you provide. While some of PennyCat users are coming to get discount coupons, others are coming to play the game. While some of them talk a lot with the bot, some of them talk less. Are the discount lovers talking a lot with the bot or gamers? Why do they stop talking at some point? So, questions can be rising in this way.
The answer is filtering conversations to find out which users tend to play the game and which users like to get discount coupons.
After filtered conversations, PennyCat dived deep to listed conversations to see the user behaviors. Below you’re seeing one of the listed conversations. In here, they easily saw the user details like first and last seen by the user, total conversations and number of min/max conversation steps. And the right side, they discovered user’s previous conversations and responses to the bot.
Enhancing the bot with understanding user behaviors was highly important for their team because the mission of PennyCat is to enable people to save their time and money on shopping. If they offer right content to the right audience with right flow, they will reach their mission.
So, after they explored user behaviors, they shaped the flow of conversations according to the needs of the right audience. They realized two things about discount lovers and game lovers;
If discount lovers start the conversation to get discounts, they continue to conversation to search new coupons rather than playing the game. But for game lovers the case is different. When game lovers who start the conversation to play a game, they can continue the conversation after their game finishes to search discount coupons.
According to this behavior, the team started to offer “Check discount” option to the users who finish their games. And they started to offer new coupon search for users who finish discount search rather than offering play game.
This boosted their average session length by 87%.