[Case Study]: How tracking metrics and iterating helped MojiHunt increase its engagement rates

There are many challenges which chatbot makers face while creating their chatbots. Of course, all the bots are different and it’s impossible to tell what the biggest challenge is for all of them. However, one of the main objectives for entertainment bot makers is to increase engagement. And also, raising the average time that user spends on playing. Did you know how easy it is to increase engagement and other metrics using analytics? Improvements to these important metrics are possible thanks to chatbot analytics tools.

MojiHunt is a great example of a bot which got 300K+ users. In total, it has 17 million emoji messages in 2 months with 17 minutes average session time. All thanks to deep analysis, correct conclusions and continuous iteration based on the data output regularly.

How MojiHunt was created?

Leo, founder of MojiHunt, observed that people in the subway are playing some dumb games just to pass their time on the way home. At that moment he came up with the idea of creating MojiHunt, chat-based emoji blast game. Sounds stupid? Maybe, but in a short period of time it became one of the most popular gaming bots on Facebook Messenger.

MojiHunt bot offers a game you can play anytime and it attracts a lot of traffic. But, of course, not all the users will engage with the bot longer than a couple of minutes. That’s why Leo wanted to find a way to lengthen the average time which users spend playing and increase engagement. Are you curious how he did that?

 

 

 

 

 

 

 

 

 

 

Acquiring first 100K users with zero marketing budget

While still testing his bot idea, Leo decided to use simple marketing methods without spending lots of money on it. MojiHunt founder decided to post screencast video on a Facebook page and make the title, “The Most Stupid Game for Facebook Messenger” to decrease the expectation from users and differentiate MojiHunt from other rich media games on iOS/Android. He made the post and decided to spend 20 dollars on Facebook video promotion. It worked. Later on, he opened Facebook Pages Manager and found out that people loved it! Over 3,000 people in 6 hours clicked the link after watching this video. Then, he tried to ask players to put a comment on the video and share their video to their friends. Afterwards, he’s made 10 new levels and set the content locket.

After these two actions, 12,000 comments, 31K likes, 7.3K shares, 1.9M views occurred. At the first stage Leo tested his idea, which proved to be a great success. But that wasn’t the end. To better understand thousands of users talking with his bot and measure their engagement, Leo decided to start using chatbot analytics tool – Botanalytics. Huge audience is not enough. On the second stage of building a great chatbot, it’s crucial to analyze the data, detect bottlenecks and improve user’s experience.

Tracking metrics and analytics matter

The key to success is to find main bottlenecks and points where bot’s users are struggling by tracking the conversion on each step of chatbot. Except standard analysis tool, Leo decided to use Botanalytics Assistant, which is a virtual assistant for bot makers to get insights about their bots automatically. Assistant managed to increase engagement and find the bottlenecks on the bot by machine intelligence & examining the whole bot performance.

One of the main findings was that some users couldn’t notice emojis on the gifs, that’s why they weren’t able to pass to the next level. Moreover, Botanalytics discovered which icons are confused with each other and caused the wrong answers. Thanks to this information, bot creators found some patterns of the possible improvements in the bot and could eliminate the elements that frustrated users at most, causing them to leave the game.

Based on the reports generated by BA Assistant, MojiHunt was improved and all struggling points have been fixed. Thanks to the improvements, average session time increased from 5 mins to 17 mins. Tracking each conversation steps also helped to increase MojiHunt’s engagement rates.

Leo: “Botanalytics Assistant helps chatbot makers to make the decision process easier. It gives hints about what exactly we should look in order to improve our bot. It helped us to identify bottlenecks and negative reactions of our users on different stages.”

Now it’s time to focus on retention rates

Leo was asked what will be his next move, he replied that right now they’re focusing on the other common problem of chatbot developers, which are retention rates. He says: ‘’If you have the same content for 3–4 days, you can expect that you will lose 90% of your users shortly thereafter. So, you have to create daily content in order to keep users’ interest.’’ Chatbot makers should be observing day to day retention rates. Also, they should filter specific conversations detecting why users are losing their interest or what attract them in particular. Leo expects that MojiHunt will prove its retention rates within a short time with the solutions that chatbot analytics provides. Also, he currently focuses on this goal.