While many have already switched to OpenClaw / Claude Code and other agents that do work for employees, I'm late and coming to what people did last year - learning to generate organic traffic on my own.

💡The idea is simple: without an advertising budget, with the help of contractors (or AI agents) just make content that brings people in by itself. It's definitely slower than buying ads, but cheaper and more effective in the long run.


So with the team we decided to try launching a "content factory" format (buzzword of last year) to promote our app among regular players. We prepared 30-40 short videos, uploaded them to Instagram, and started watching what happens.

Video format: for admins, humor, aesthetics

In 7 days - more than 1k live subscribers and 600k total views. The entire result was pulled by one video that got into recommendations with 427k views. The other 17 videos just exist nearby, gradually gaining views. All videos have our banners 🖥

The audience turned out to be very good: Russia, 25-34 years old, men and women equally.

On other platforms: YouTube - zero impressions, TikTok not tried (only CIS views are relevant there), VK Video we don't believe in. Instagram is the only one with any response so far. We focused on it.

❗️What doesn't work - app installs. There are very few (15 installs during experiments). People watch, subscribe, but don't click. This is the next skill to improve - we don't yet know how to manage traffic from views. When we learn, I'll come back and tell you happily.

Attempts: we tried uploading videos from different accounts, watched how it affects reach - with FB connected it works better. Now we're testing two more hypotheses in parallel - including making a digital twin of our marketer who will talk about Telegram, bots, and marketing in general.

What will come of this - I don't know yet. But we're actively testing, and when there are results, I'll share. For now I can say that hypothesis testing is very expensive 😅