Some teething problems:
- Bot detection -> have to run headful puppeteer on docker which took a few trys
- Queue mitigration had a few bugs
- Some modules were broken due to site changes / anti bot (happens to be the most profitable module, would be expecting ~$200-500/month if all uploads were...
Time to scale
There have been no bans, warnings. So it's safe to create and run more accounts without setting up a mobile proxy farm.
Technical steps taken:
- Change queue system to global queue system, before each account had their own queue - now there is one global queue for all accounts
Potentially, but emulators are heavy on ram, so you would struggle to run more than one active session on a RPI (8GB) - and it defeats the purpose of using RPI's, instead just buy a beefy server.
Also emulators have their own fingerprints
Figure out how your going to get traffic from the platform of your choice.
Record (video, url, written description etc) each step of interacting with that platform to get the desired result, logging in, posting etc etc.
Does anyone know how to see all your metrics in google analytics combined into one across multiple properties?
There's google data studo - but that has a limit of 10 properties.
There's EKM services - They charge $150-500 minimum per month
Google analytics 360 - Native rollup dashboards, only...
Was on a call with someone who works at the competitor to site 5, and I was told that I wasn't skilled enough, so you know the drill.
- Scraped competitor's content + other data sources in the niche
- Built a custom designed site to better represent the data
- Use some AI magic to...
Scale horizontally, so look for a hosting provider that provides 0.5GB ram 1 core servers that has API access (Azure, GCP, AWS, Linode, etc)
Optimize your code, one browser tab can be isolated cookies / session storage wise - so run 10 per browser to save RAM / CPU cycles.