5 Fun facts about 100.000 users chatting with a cat

Mica with the Facebook Messenger QR code

After a tough and exciting 10 months creating and developing a cross-platform chatbot (chatcat :) ) we finally reached one spectacular goal: 100.000 users!

YAY! 100.000 chatbot users!

Mica, the Hipster Cat Bot is a chatbot, a service integrated in messenger apps, that helps you discover the best places worldwide. Mica was one of the very first chatbots on the Facebook Messenger and Skype platform — and is still the top 7 bot on Skype.

The original Mica, the Hipster Cat

To celebrate this great number of 100.000 unique users we want to give you an insight into our precious secret usage statistics — and share five fun facts about our project:

Fun fact #1: People don’t chat during work

People like to chat on the weekends or in the night and tend not to use messenger apps during work. At least only a few people chat with Mica during work hours, although a helpful purr would increase their productivity significantly.

Users by time of day — Google Analytics stats

Wednesday and Friday night are particularly low, but on Saturday and Sunday afternoon we have the most traffic.

Fun fact #2: Telegram users don’t like cats

Since we have been featured on Microsoft Botframework we receive most of our requests from the Skype platform:

Total number of users per platform

The Telegram integration of Mica is the one with the least traction. In total we have only around 1000 users. One interpretation might be that the platform is not that widely used, another that Telegram users are more into dogs.

We also have a few users on Kik, but since we launched on Facebook and Skype much earlier, the total numbers of Kik users is relatively low — That said, it is our fastest growing channel.

The Hipster Cat Bot implementations on other channels such as WeChat, Line, Slack and Viber are still in Beta — Ping me if you want to check out an early version of Mica on these platforms.

Fun fact #3: People love cat pictures

So far Mica had more than 500.000 incoming requests, that divides into approx. 5 requests on average per user. From these incoming requests 130.000 resulted in venue requests (350.000 venues where displayed in total, since some platforms support carousel lists).

Analysis of user input

And what do the others want? Cat pictures, of course! And one is not enough: After a user receives a cat picture or animated GIF they usually request another!

Fun fact #4: Mica is huuuuge on the Philippines, like really!

Some interesting demographic data:

Users by country (all channels)
Users by language

When we looked at the users’ languages, we didn’t see anything unexpected: English is dominant since Mica communicates in English. After the United States and the UK the next countries are Austria and Germany, because our company’s friends and startup community is based here and Mica received a lot of PR in these German speaking countries. So no big surprise here.

But when we have a particular look at the Facebook users we see something funny:

Facebook users by timezone

There is a peak in timezone 8! And after we analyzed the data we realized, that many users request venues in the Philippines! (Despite the fact, that we never targeted that particular region, or any other region concerning marketing whatsoever).

While we love to attract people in Central Europe, it’s also nice to see that Asia has a heart for talking cats. And it’s nice to see that our international service also attracts international users.

Fun fact #5: 15% come back the next week

One of the most important metrics when you have a (B2C) startup are the cohorts: Do the users come back after using the service?

Typically you can say that a website or app has a good retention when you have 20–30% returning users (daily or weekly, dependent on your service). Chatbots typically have bad retention, because of two reasons: Most chatbots are designed badly and don’t offer a good and welcoming UX or an engaging dialog. The second factor is that many people just want to try out chatbots without really being interested in the service simply because they are new. So a big part of the users are chatbot newbies that simply are curious to see how a bot or better said a dialog with a bot works.

With a couple of thousand daily active Mica users we see, that users generally don’t use the service daily, but rather weekly. The weekly cohorts are much better than the daily which also reflects the type of service we offer — looking for a hip venue is typically not a service you use every day, but every couple of days.

Most important statistic chart: Cohorts (Google Analytics)

Another interesting statistic is the active users chart: In this chart you’ll see that the 30 days active users is by far the biggest group. So, users love to come back and talk to Mica.

Active Users Statistic in Google Analytics

Meow me!

Creating Mica, the Hipster Cat Bot was a great challenge. We had to deal with more complicated topics such as data security and privacy in chatbots, some of which we did not anticipate. Also the difficult, almost fragile, use of emoji in chatbots is a UX topic, we could spend hours talking about.

However, it was super interesting and exciting to create a virtual personality people love to chat with!

We would love to get your feedback on our statistics! What challenges did you face creating or chatting with a bot? Leave us a Meow!

Interested in more about bots? Join us at the BotBarcamp.wien end of April in Vienna ❤

Senior Software Engineer, Public Speaker, Founder of Women And Code