Lately, everyone’s talking about different Telegram bots on every platform. So why not? I’ve got something to share too.
🙀 What we've done?
We created a web bot where you can fill out a profile and then swipe through cards with info about other users, giving them likes or dislikes. And if the interest is mutual, you can start chatting.
Sound familiar? Yep, we basically put Tinder in Telegram. Not the whole thing, of course, but the main flow.
Sound familiar? Yep, we basically put Tinder in Telegram. Not the whole thing, of course, but the main flow.
Now this Tinder-like feature lives in various brand communities, blogs, business clubs, and so on. Members use it to network and find people with similar interests.
🚜 How did we end up here?
Back in 2020, when everyone was hiding at home from COVID, all offline events were canceled, and event organizers were pulling their hair out in panic, we (just one manager and one full-stack developer) dove headfirst into the world of conferences with our online event format.
I won’t dwell too much on the story of how two guys managed to work with all the major enterprises in Russia within a year, but it was a wild ride.
While talking to new clients (the organizers of these very conferences and forums), we discovered that people attended not so much for the smart lectures and coffee breaks (although you’d think so) but for valuable networking opportunities.
While talking to new clients (the organizers of these very conferences and forums), we discovered that people attended not so much for the smart lectures and coffee breaks (although you’d think so) but for valuable networking opportunities.
It became clear that online streams and live interactive sessions, which we were offering, were great, but we needed to figure out how to connect people in an online format.
Sure, by that time, there were already things like random coffee chats and participant directories. But we couldn’t just steal such a simple solution. So, we “stole” (like artists, of course) a more complex one—the interface of a dating app.
We built the first version on the web. In that version, our “business Tinder” was just one of the tabs on our online event platform. Online forum attendees would watch the live stream while swiping other viewers' profiles left and right, building new professional connections. Organizers were thrilled, users were happy, and we weren’t complaining either.
There was one major downside to the web format: the day after the conference, site traffic would drop to nearly zero, making it harder to get a mutual like.
On top of that, it was difficult to notify users about new matches or messages through the web. Sure, there are web push notifications, but how many of you actually allow them?
On top of that, it was difficult to notify users about new matches or messages through the web. Sure, there are web push notifications, but how many of you actually allow them?
🥾 The kickstarter moment
At one point, we were creating a platform for the "Colleagues" conference for the Setters agency—awesome folks in marketing and creativity. When I offered them our web solution for networking, their CEO, Zhenya Davydov, asked if we could implement a "Tinder" as a web bot for Telegram.
That question fundamentally changed the direction of the product. Telegram turned out to be the perfect place for hosting a networking bot.
That question fundamentally changed the direction of the product. Telegram turned out to be the perfect place for hosting a networking bot.
🐞 Bug or feature?
While observing how people use our "business Tinder," we noticed an interesting pattern: participants in the online format mindlessly like every profile they see, without even reading the descriptions, just hoping for a mutual like.
Sure, this strategy worked well for collecting as many matches as possible, especially when there were plenty of these "active" users. But as our stats on initiated conversations showed, very few people were actually ready to engage with that many new connections.
The result? Hundreds of new contacts in a day, most of which were completely irrelevant. Based on customer development research, the most productive new connections were made by those who received 5-10 matches per day.
To naturally encourage users to actually read profiles and like only genuinely interesting participants, we limited the number of likes to 10 per day. This change definitely didn’t sit well with some users who loved liking everything that moved, but the conversion rate to initiated conversations skyrocketed.
The result? Hundreds of new contacts in a day, most of which were completely irrelevant. Based on customer development research, the most productive new connections were made by those who received 5-10 matches per day.
To naturally encourage users to actually read profiles and like only genuinely interesting participants, we limited the number of likes to 10 per day. This change definitely didn’t sit well with some users who loved liking everything that moved, but the conversion rate to initiated conversations skyrocketed.
📱 Why not an app?
This is probably the most frequently asked question by clients, so I’ll answer it before the comments section gets flooded.
User Base 🌎
Telegram has 700 million active users, and the effort required to attract users into a bot is significantly lower than dragging them into a separate app.
Works on All Platforms 📟
Sure, you could invest more resources and develop a native app for iOS and Android. But then you'd still get complaints like, "What about my Huawei?" or "I need it on desktop," or "You forgot about my Windows Phone." Thanks to Durov, anyone who can launch Telegram can use our service. And if they can’t—well, they can take it up with Pavel about their Huawei.
Fast Launch 🚀
No store moderation or filling out tons of app and company information. We can build a new bot in about an hour now. And soon, we’re planning to roll out an admin panel so we won’t have to do anything manually—users can configure everything themselves.
Plus, cool features like sending messages on behalf of the bot, no need for installation, and low launch costs make the web bot format just awesome.
Plus, cool features like sending messages on behalf of the bot, no need for installation, and low launch costs make the web bot format just awesome.
🎯 What's Next
There’s still work to do, so (mainly for my own motivation) here’s a public roadmap of what we plan to improve next:
Figure out the filtering by categories
Right now, we only allow filtering profiles by one group of tags. If there are more groups, the system starts to creak and everything works painfully slowly. So yeah, we need to optimize this!
Add a smart recommendation system
We’re (not very actively yet) working on having AI analyze user profiles and prioritize showing the most similar people by interests.
SaaS
For now, we build each new bot from scratch every time. In the near future, we plan to finish the admin panel and completely shift the launch tasks onto the users themselves.
🙅♀️ Outro
’m open to comments from anyone who cares. So, if you have feature suggestions for our backlog, feel free to drop them here in the comments or DM me on Telegram.
And if you have experience with other similar services (like random coffee, etc.), I’d appreciate hearing what you liked or what annoyed you about them.
And if you have experience with other similar services (like random coffee, etc.), I’d appreciate hearing what you liked or what annoyed you about them.