Picabu Effect Analysis

After launching the application, the question always arises of where to get users. In this article I would like to share my experience on how, at the start of the project, without cash investments, I got a good starting traffic.

Bots in instant messengers are gaining popularity. On this wave, there was a desire to get acquainted with the telegram api. As HelloWorld, I decided to write a bot that converts voice messages to text. HelloWorld turned out to be pretty working, and this time I decided not to bury the project in the bowels of the github, but to launch it in production. Feel like a startup.

Launch the project was not work. So it works, but what to do next? The bot is used by two people, and even those in the role of testers, since among my contacts there were no fans of voice messages. So how to tell the world about him?

In the beginning, I wanted to try free user search methods. And the first thing that my choice fell on was the pickup resource. For those who are not up to date, Pikabu works on the principle of user-generated content. Any user can create a post on any topic. The post falls into the “fresh”, where quite quickly it is lost among the posts of other users. But if a post in a short time collects a lot of positive ratings, then it will be on the main page “hot”, and here it is already guaranteed thousands of views.

Getting into the hot is not so easy, and yet, I succeeded.

So, on September 1, I decided to tell the community about my bot. I wrote down gifs in which I visually showed how to use the bot, wrote a short description, and the post went fresh. After a couple of minutes, the first comments had already begun to appear, and instead of the post rating, the watch icon was shown. After about an hour, a notification came from pickab that the post had gotten hot. In the first 4 hours of publishing on picabu, the bot was added to 800 chats, and 2,300 voice messages were processed.

I was curious how long the influx of users would last, how fast the number of chats in which the bots added increased, and, most importantly, how many people would continue to use the bot after the post crashed out of hot. And so, after three weeks, I decided to conduct an analysis of what it is, “The Picabu Effect”.

The post was published on September 1 at 18:00. At the beginning, the post was actively gaining a rating, it lasted about a day and then the growth slowed down sharply. The second day, the rating continued to grow, but growth was much slower. At the end of the second day, the growth froze. Graph for clarity:

Active rating growth lasted about a day and changed from 0 to 3000. Then the growth slowed down sharply, and over the next day changed from 3000 to 3400. Over the next two weeks, the rating changed from 3400 to 3500, that is, we can say that the growth has stopped.

Returning to the bot: how did the increase in the rating of the post on the picaba affected the appearance of users and the number of voice messages sent to the bot? There are no surprises or surprises here, the growth of bot metrics is completely synchronized with the growth of the post rating. For clarity, the graph:

The graph shows that users are increasing at the same speed with which the post is gaining a rating. But the growth of messages is jerky. If you look at the time, you can see that the active growth of messages goes during the day and slows down at night, which is logical in principle, people communicate more during the day and less at night. Additionally, you can verify this if you build a graph showing how many new chats the bot is added, and how many voice messages are processed per hour:

The timeline shows the most active growth of users and messages in the daytime, and growth slows down at night. The graph shows many sharp jumps in message growth. What is the reason for this - I do not know. But there is an assumption that it was either someone actively checking the quality of work, or the bot was added to chats with a large number of voice messages, but after a while it was deleted.

Well, in the end, the most curious metric. Two weeks after the publication of the post on pikabu, after all the hype subsided and the post long ago left hot, how many people really began to use it?

The most surprising thing for me is that until now new chats with a bot continue to appear, not a lot, about 25-30 new chats a day, but nonetheless. I don’t know where they come from, maybe they somehow find a post on a picaba, see a bot in public chats and add them to their conversations, or maybe “word of mouth” just works. On the day, the bot continues to process a little less than 1000 voice messages, which is good news.



Well, in conclusion, what conclusion can be made? Picabu can give a lot of traffic to new users, but for a short period. However, about 10% of these users will remain with you. But still, getting into the hot is not such an easy task, here I think that I was just a little lucky in the sense that the topic of voice messages is quite actively discussed by the community, and the post did not go unnoticed.