Olwi
Reddit marketing tool
It's hard to find potential customers and get engaged in the right discussions
It's been a while since the latest update. I skipped IH for a while and worked on the product.
A few months ago I decided to pivot: from a tool that helps to find potential problems to solve to a tool that helps to find whatever you want on Reddit. I decided to follow the bigger potential.
I didn't remove the searching for problems completely: one can select it as a purpose of research and the tool assembles the necessary search query.
I've created a blog to share how to do something on Reddit, e.g. market a project, find relevant threads to engage with.
I've created a bunch of landing pages for various target customers. I'm trying to optimize many keywords for SEO.
Recently, I created a post on IH and got around 30 new users.
It sounds like a small milestone, but there was a ton of work to do this. Initially, I used ready-made packages to handle text classification. They were ok, but they also classified irrelevant data.
I decided to build my own neural net. I didn't know ML. I went with TensorFlow.js. Tried to build one for 3 months. Some of them were ok, but also missed a lot. Then, after learnings, I went with a pre-trained model.
Eventually, I'm using it.
Other updates:
- 180 newsletter subscribers. I'll resume hand-picking new problems to solve and sending them to people.
- Redesigned a bit 2 UI components.
- Fixed issues regarding deployment. So olwi.xyz is up and running.
3 days since the newsletter launch and 52 people subscribed so far. I created a few posts on Reddit with the hand-picked problems to solve from Olwi. The Reddit community is huge, some people got subscribed. Currently, I don't know how frequently I'll be sending the newsletter(once a week maybe), we'll see.
I decided it's time to create a newsletter where I might share the hand-picked and latest problems from Reddit. Created a Mailchimp account, set up a campaign. Bought a domain - olwi.xyz and a professional email.
I'll be collecting subscribers. The newsletter
Configured the domain, deployment and decided to show how these things are going. I worked on the product for the past several months and it's hard to understand when it's sufficient to show it because of its specifics: it uses a neural network that I constantly train on new data. But here is a hell of a lot of it.
I decided to receive any feedback on a stage where the model started to give relevant data .