Every hosted platform that runs your billing takes a slice of what you earn. The slice feels small when your revenue is small. Then your revenue grows, the slice grows with it, and you notice you are paying more for the same software than you did a year ago while the software itself has not changed. This is the quiet cost structure that most agencies and SaaS operators sign up for without doing the arithmetic. This piece does the arithmetic.
A percentage cut has one property that makes it expensive: it scales with your success and never stops. It is not a fee for a fixed amount of work the platform does. It is a claim on a fixed share of everything you sell, forever, regardless of whether serving your thousandth customer costs the platform any more than serving your first.
Look at what the common platforms charge, as reported and approximate:
The first three are pure percentage models. The last is a flat subscription, which behaves very differently as you scale, and we will come back to why that distinction matters. On top of any of these you still pay your payment processor its own fee. That processor fee is unavoidable. You pay Stripe or its equivalent whether you self-host or not, so it is not part of the comparison. The platform cut is the part that is optional, and it is the part worth examining.
Self-hosting on VBWD replaces the percentage with a fixed server bill. VBWD runs on a small EU VPS: Python/Flask, PostgreSQL, Redis, and a Vue 3 front end, all in Docker. A box that runs this comfortably costs around €12–16 per month. That number does not move when your sales go up. It is the same bill at €1,000 of monthly sales and at €50,000.
Put the two models side by side. Assume a 10% platform cut, which is the going rate for the subscription and digital-goods platforms above. Ignore the processor fee on both sides, because it is identical on both sides.
Monthly sales 10% platform cut VBWD server (fixed) Difference/mo
------------- ---------------- ------------------- -------------
€500 €50 ~€14 ≈ €36
€1,000 €100 ~€14 ≈ €86
€5,000 €500 ~€14 ≈ €486
€20,000 €2,000 ~€14 ≈ €1,986
€50,000 €5,000 ~€14 ≈ €4,986At €1,000 in monthly sales, a 10% platform takes €100 every month and keeps taking it. The server bill stays where it was. At €20,000 the platform is taking €2,000 a month for software that costs it no more to run than it did when you were paying €50. The gap between the two columns is not a rounding error. It is the majority of the cost, and it widens every time you grow.
The flat-subscription model, like GoHighLevel's tiers, sits between the two. A fixed monthly platform fee also stops scaling with revenue once you are over its threshold, which is better than a percentage. But it starts high. At $497 a month you are paying roughly thirty times the VBWD server bill before you have sold anything, and you still do not own the stack underneath it. You can see how VBWD's own commercial terms compare on the pricing page: the software is source-available under BSL 1.1 and free for commercial use until your VBWD-attributable annual sales exceed the value of 6.7 BTC, after which it converts to Apache-2.0. Below that ceiling your only recurring software cost is the server.
The break-even is easy to find and it arrives early. The platform cut equals the server bill when 10% of your sales equals about €14, which happens at roughly €140 in monthly sales. Below that, the percentage platform is cheaper. Above it, self-hosting is cheaper, and the advantage grows without limit.
Read that again, because it is the whole argument in one line: the crossover happens at around €140 per month of revenue. Almost any serious operation clears that in its first weeks. Past the crossover, every additional euro of revenue that a percentage platform would have taxed at 10% stays with you, minus a server bill that does not care how much you sold.
This is why the percentage model is comfortable for platforms and expensive for you. It is cheapest exactly when you have the least to lose, and most expensive exactly when you can least ignore it. A fixed cost has the opposite shape. It is a larger share of a small business and a rounding error for a large one.
The cost math is the loudest reason to own your stack, but it is not the only one. When you self-host, your customer list, your content, and your billing relationship live on a server you control rather than inside a vendor's account. That has three practical consequences.
VBWD is built so that this ownership is real at the code level, not just a marketing claim. The core is deliberately agnostic and everything domain-specific lives in plugins you can inspect and swap; you can see how that is organised in the architecture overview and browse the available plugins. The billing engine, the subscription lifecycle, and the invoicing all run inside your own deployment rather than behind a vendor's API.
None of the above is free of work, and pretending otherwise would be dishonest. When you own the stack, you own everything the hosted platform was quietly doing for you:
These tasks are well-understood and VBWD ships to make them routine: Docker for reproducible deployments, PostgreSQL and Redis as standard components with standard backup tooling, and operational guidance in the documentation. But "routine" still means someone does it. Budget for a few hours a month of maintenance, or the cost of a managed host or a contractor who handles it. Fold that into the comparison honestly. Even at a generous estimate for your own time, the numbers in the table above leave a wide margin once you are past a few thousand euros in monthly sales.
And there is an honest version of the other choice. If you would genuinely rather pay a percentage in exchange for never thinking about a server, backups, or a certificate renewal, then a hosted platform is the correct decision for you, not a mistake. The percentage is the price of not owning the operational burden. For a small side project, or for an operator whose time is worth far more spent elsewhere, that trade can be the right one. The point is to make it deliberately, with the arithmetic in front of you, rather than by default.
The decision comes down to which cost curve you want to live on. A percentage cut is a variable cost that grows with every euro you earn and never stops. A server bill is a fixed cost that a growing business outgrows in the opposite direction, becoming a smaller share of revenue every month. The crossover between the two arrives early, at roughly €140 per month of sales, and past it the fixed-cost model wins by a margin that widens indefinitely.
What you buy by paying the percentage is the freedom to not run infrastructure. What you buy by self-hosting is margin that compounds, data you control, and insulation from a vendor changing the terms underneath you, in exchange for owning uptime, backups, and patching. Run the numbers for your own revenue, add an honest estimate for the maintenance you will actually do, and the answer usually stops being ambiguous somewhere in the low thousands of euros per month. You can review what runs on your own box and what stays yours on the features overview and the billing documentation.
Next: work through the numbers for your own operation with the cost breakdown on the pricing page, then read the deployment guide in the docs to see exactly what owning the stack asks of you.