← Back to library
Basics Theory

The economics of a VPN service: where the margin comes from

A service is killed not by blocking but by bad economics. Let's break down what per-client cost is made of, where the margin comes from, and what people most often lose money on. Doing the unit math is a skill without which growth turns into loss.

This material covers the engineering of your own network infrastructure and is educational in nature. Keep the money above board: the service being publicly anonymous doesn't make you invisible to the bank and the tax office.

Count by a single client

The first thing to internalize: a service is killed not by the regulator or by traffic inspection but by bad unit economics — expensive idle servers and low customer return. You'll survive a block if reserves are set up. But a quiet loss on every client piles up unnoticed and one day buries the business.

The unit is a single client. All the economics is counted through them: how much they cost you per month and how much they bring in. If they bring in more than they cost, you're in the black and can grow. If it's the other way around, every new client deepens the hole, and growth won't save you — it'll speed up the fall.

What the cost is made of

A service's expenses aren't "a server for 5 euros." The full picture by line item:

  • Exit-node VPSes — a few euros a month each. But one node handles dozens to hundreds of clients, so per client it's pennies.
  • Panel and bot VPS — one machine for the whole service, also a few euros.
  • In-country relays — if you do cascades, cheap Russian servers for entry are added.
  • Domains — from pennies to a couple of euros a year each.
  • Payment-processor fee — a few percent of every payment. This is a variable line item, growing with turnover.
  • Sourcing white IPs — nearly free per address, but needed constantly.

Per-client cost is figured like this: (sum of all VPSes + domains + relays per month) divided by the number of active clients, plus the processor fee on their payment. The key takeaway: the more clients on a single node, the lower the cost of each. The infrastructure is nearly fixed, and you can hang many clients on it — so scale works for you.

Where the margin actually lives

The beginner's main mistake is thinking the money is in the one-time sale. It isn't. The margin is made on retention.

  • Annual plans are cheaper per month but give money up front and sharply reduce churn. A client who paid for a year doesn't think every month "should I unsubscribe?"
  • Count LTV against CAC — how much a client brings in over their whole lifetime versus how much it cost to acquire them. LTV should be many times higher. If you spend more on acquisition than a client brings in over their lifetime, you're buying a loss.
  • Don't undercut to zero. Too cheap — abuse piles in, load grows, and there's no money. The cheap client is often the most problematic.

Plans as a tool, not a price list

The plan grid isn't "pick a price" but a way to sort your audience by wallet and needs. The working logic:

  • Trial — 1–3 days with reduced traffic. The job is to lure them in but not let them abuse it.
  • Basic — one node, a couple of devices. A cheap entry for those who are sizing you up.
  • Main — all nodes and auto-select. The mass plan; this is where volume is made.
  • Premium — everything plus beating "jammers" and priority. Higher price, higher margin, for the demanding.

Technically, plans are sets of access rights in the panel. A client upgrade isn't a "re-registration" but adding them to the right set. More on this in detail in the "Panel" section.

What kills profit

Four main killers, each of them unnoticeable until it's too late:

  1. Downtime during blocks. The client couldn't connect — left and posted a negative review. Treated with reserves (cascade, CDN, spare IPs) and fast address swaps.
  2. Abandoned paid servers. A stopped or unneeded VPS keeps ticking on the bill. Delete the excess, monitor balances — this is the dumbest and most common leak.
  3. Silence in support during a mass outage. Number one cause of churn. A client will forgive a breakage but won't forgive silence. A reply template and a status in the channel save your reputation.
  4. Trial abuse and key sharing. They eat resources without payment. One account on ten devices is money lost for nothing.

A mini-formula to keep in mind

If you boil it all down to one line:

Monthly profit ≈ (Active clients × Average check) − (Servers + domains + relays) − (Fees) − (Your support time).

This formula grows via three levers: higher retention (annual plans, fast support), lower downtime (reserves against blocks), fewer dead servers (infrastructure hygiene). Everything else is detail of these three levers.

We'll dig into the economics more deeply in the "Business" section — there it's referrals, the funnel, and retention in numbers. For now hold onto the main thing: you have to count the unit from the first client, not when it already hurts. A service that doesn't know its own cost runs at a loss and doesn't suspect it.

Next guide Glossary: VLESS, Reality, SNI, DPI, node, inbound in plain words → Article unclear or something off? Message me and I will help or fix it. @notrealvpn →
This material is educational and covers network-infrastructure engineering. You are responsible for complying with the laws of your jurisdiction.