← Back to Blog

Month Two Running a VPS Hosting Business: What Actually Changed

Published May 3, 2026  ·  6 min read  ·  Galaxy Cloud Solutions

The first post covered the setup — the hardware, the software stack, why someone would start a VPS hosting company out of a home in Valley City, North Dakota. This one covers what happened once real people started signing up and paying money.

Short version: the infrastructure held up. The humans were more complicated.

The PayPal Dispute

The most stressful thing from the past few weeks was a PayPal dispute from the first paying customer. They had been running bandwidth-reselling apps in the background since day one — Honeygain, EarnApp, and a few others — which the AUP explicitly prohibits. The hourly compliance scanner caught them, I suspended the VM and sent the violation notice, and within about 20 minutes there was a dispute filed.

How it resolved: everything was documented. The AUP is clear, the suspension notice is timestamped, and the compliance script logs exactly what it found and when. PayPal reviewed it and found in my favor. Seller Protection covered the chargeback.

The lesson that stuck: automated enforcement with clear logging is not optional. If I had suspended that VM manually on instinct with nothing documented, that dispute likely goes the other way.

What the Monitoring Actually Caught

I built out a lot of monitoring scripts before launch and honestly was not sure how much of it would matter at this scale. It earned its keep. Beyond the AUP violation, the brute force monitor has caught and logged several SSH scanning attempts that fail2ban handled automatically. Bandwidth alerts fired on a couple of traffic spikes I could notify customers about before they hit their limits. The availability monitor caught one brief VM hiccup and auto-restarted it before anyone noticed.

There are scripts I have not needed yet — disk I/O limits, connection rate thresholds, things built for abuse scenarios that have not happened. They are there when they do.

Blog Traffic

The AWS Lightsail comparison is the clear winner so far — picked up search rankings fast and drives steady daily visits. The FiveM tutorial does well with a specific audience that converts reasonably. Personal story posts like this one get fewer visitors but the time-on-page is much higher, which means people are actually reading rather than bouncing immediately.

Organic traffic is modest but growing week over week. The Google Ads campaign is at $5/day, which right now is about figuring out what messaging resonates rather than driving volume before conversion rate is proven.

What Changed in the Product

The customer dashboard went from functional to genuinely useful. At launch, customers could see VM stats and reboot. Now they can reinstall the OS entirely from the portal, manage automatic snapshot schedules, reset their root password, and get into a browser-based SSH console without needing a local client. The console feature took a while to get right end-to-end, but it handles a huge portion of what would otherwise be support tickets from people locked out or on a machine without their SSH keys.

What Is Not Working Yet

Word of mouth has not started compounding. The customer base is still too small for referrals to matter much. Directory listings drove short traffic spikes that did not convert heavily — Product Hunt, Hacker News, HostAdvice. Useful for backlinks, less useful for direct signups this early.

Fake account signups happen more often than expected. People registering with garbage email addresses and never paying, apparently just probing to see how the system responds. The auto-suspend and cancel logic handles it without manual intervention, but it is a minor annoyance to see in the dashboard.

Month Three

More content. Competitor comparisons rank well and reach people actively evaluating options. Tutorials for developers and game server operators bring in audiences with specific, concrete needs who tend to stick around once they find something that works for them.

The server has capacity. The stack is stable. The main job right now is making sure enough people find the service and that their first experience is good enough they renew.

If you are running something similar or thinking about it, feel free to reach out: [email protected].