Ownership & economics
You own your stack: self-hosted vs platform fees, the real 5-year math
Renting a platform is cheap in month one and expensive by year five. The real answer is that platforms win on speed-to-launch and hands-off simplicity, while owning an open-source stack wins on long-run cost, control, and the value of the asset you keep. This guide runs the actual five-year math so you can see the crossover point instead of guessing at it.
On this page
The two ways to pay for a website
Every website is paid for in one of two shapes. You either rent — a monthly subscription to a platform that hosts your site, owns the code, and often takes a cut of your sales — or you own — a build on open-source software that runs on infrastructure you control, where the recurring cost is just the server and your own time. The rent shape front-loads convenience and back-loads cost. The own shape front-loads build effort and back-loads value.
Neither is universally right. The mistake is comparing them on month-one price, where renting always looks cheaper, instead of on the full horizon you actually keep a website — which for most businesses is three to seven years.
A worked 5-year model
Here is a concrete, illustrative worked example — not a survey figure, just transparent arithmetic you can re-run with your own numbers. Take a small store doing roughly $12,000/month in online sales ($144k/year). Compare a mid-tier hosted platform plan against a self-hosted open-source build (Next.js storefront + Medusa, or WordPress + WooCommerce), hosted on a managed server.
| Cost line | Hosted platform (rent) | Self-hosted open-source (own) |
|---|---|---|
| Upfront build | $0–2,000 (theme setup) | $8,000–15,000 (custom build) |
| Platform subscription | ~$2,400/yr ($200/mo tier + apps) | $0 |
| Per-sale platform fee (~2% of $144k) | ~$2,880/yr | $0 |
| Managed hosting | included | ~$600–1,200/yr |
| Paid apps / plugins | ~$1,200/yr (subscriptions) | ~$0–300/yr (mostly open-source) |
| Year-1 total | ~$6,480–8,880 | ~$8,600–16,500 |
| Years 2–5 (each) | ~$6,480/yr | ~$900–1,500/yr |
| 5-year total | ~$28,400–34,400 | ~$12,200–22,500 |
Illustrative worked example. Platform transaction fee assumed at ~2% of sales (above payment-processor fees, which both models pay). Your real numbers will differ — the point is the shape of the curve.
The pattern is the whole point: renting starts cheaper and gets more expensive every year your sales grow, because a chunk of the cost is a percentage of your revenue. Owning costs more to stand up, then flattens to little more than a server bill. Somewhere in years two to three, the cumulative lines cross — and after that, the gap widens in favor of owning, faster the more you sell.
The single most under-priced line on a platform is the per-sale fee, because it scales with your success. A 2% skim on $144k/year is $2,880 this year — and if you double sales, it doubles too. You are effectively funding the platform’s roadmap out of your margin. See our headless e-commerce approach for how self-hosted Medusa removes that skim.
The fees platforms don’t put on the pricing page
The subscription tier is the real number. The costs that surprise people are the ones stacked around it:
- Per-transaction platform fees — separate from, and on top of, your payment processor’s rate. Waived only if you use the platform’s own payment product.
- App/plugin subscriptions — reviews, upsells, bulk editing, advanced shipping. Each is a few dollars a month; a real store runs five to fifteen of them.
- Tier jumps — the feature you need (checkout customization, staff accounts, API access) often lives one plan up.
- Migration cost at exit — the price you pay later to leave, because your content, theme, and customer data are shaped to the platform, not portable by default.
That last one is the real lock-in. A rented site isn’t just a monthly cost — it’s a monthly cost you can’t easily stop paying, because leaving means rebuilding. Owning inverts that: the build cost is paid once, and the exit cost is near zero because the code and data are already yours. This is the same logic behind our stance on self-hosted managed hosting and it’s why we build for e-commerce and retail brands who plan to still be selling in five years.
Where the crossover actually is
Three variables move the crossover point, and you can reason about them without a spreadsheet:
- Sales volume. The higher your revenue, the sooner owning wins, because per-sale fees dominate. A store doing $500/month may never cross; a store doing $30k/month crosses in the first year.
- How long you’ll keep the site. Under ~18 months, renting usually wins on pure cash. Past three years, owning almost always does.
- How much you customize. If a stock theme genuinely fits, renting is cheap. The more you fight the platform’s limits with paid apps and workarounds, the faster owning pays off.
| Monthly online sales | Roughly when owning gets cheaper |
|---|---|
| Under $2,000 | Often never — rent and stay simple |
| $2,000–8,000 | Around year 2–3 |
| $8,000–25,000 | Around year 1–2 |
| Over $25,000 | Inside year 1 |
Rule-of-thumb crossover, illustrative. Assumes a keep-it horizon of 5 years.
When a platform is genuinely the right call
We’ll say this plainly because most agencies won’t: sometimes you should just rent. A hosted platform is the right choice when:
- You’re validating an idea and need to be live this week, not this month.
- Your sales are low enough that per-sale fees are noise, and a stock theme fits without a pile of paid apps.
- You have no technical support at all and want a single vendor who handles literally everything, and you accept paying a premium for that.
- You expect to pivot or close inside 18 months, so long-run asset value doesn’t matter to you.
When owning your stack wins
Owning is the right call once the website is a durable business asset rather than a disposable experiment:
- You’re past validation and plan to run this site for years — the horizon where the cost lines cross.
- Your sales volume makes per-sale platform fees a real number you’d rather keep.
- You need customization the platform charges for or forbids — bespoke checkout, unusual catalog logic, deep integrations.
- You care about owning your customer data and not routing it through a third party’s checkout.
If that’s you, the path is a hand-built site on portable open-source software — web design & development on Astro, Next.js, or WordPress — hosted on infrastructure you control, with a real headless commerce engine if you sell online, and automation wiring it to your other tools. It costs more to stand up and pays you back every month after. When you’re ready to run your own numbers, start a project and we’ll scope it clearly — including telling you if renting is the smarter move for where you are right now.
Questions
FAQ
Is self-hosting always cheaper than Shopify or Squarespace?
No. Renting a platform is usually cheaper in the first 12–18 months and for very low-volume stores, because the upfront build cost is near zero. Self-hosting an open-source stack gets cheaper over a multi-year horizon and at higher sales volume, because it removes the monthly subscription and the per-sale platform fee. The crossover typically lands somewhere in years one to three depending on your revenue.
What is the “per-sale platform fee” and does self-hosting avoid it?
Many hosted platforms charge a transaction fee on every order — separate from and on top of your payment processor’s rate — unless you use the platform’s own payment product. On a self-hosted open-source store the only cut is your payment processor’s standard rate; there is no platform skimming a percentage of each sale.
Do I really “own” a self-hosted site, and can I leave later?
Yes. The site is real code in your repository running on hosting you control, built on open-source software. There is no subscription that, if cancelled, degrades or deletes your site, and no proprietary lock-in that makes leaving expensive. Portability is the entire point of owning your stack.
How much does an owned build cost upfront versus a platform theme?
A hosted platform with a stock theme can be near zero to a couple thousand to set up. A custom, self-hosted build is a larger one-time cost — often several thousand up to the low five figures depending on complexity — after which the recurring cost drops to roughly a server bill. You are trading a higher upfront number for a much lower ongoing one.
When should a small business just rent a platform?
When you need to launch immediately, your sales are low enough that per-sale fees are negligible, a stock theme fits without a stack of paid apps, and you might pivot or close within about 18 months. In those cases the simplicity is worth the premium and owning would be over-engineering.
Where this applies
Industries this matters most for.
Put this into practice
We'll scope it clearly — including telling you when the simpler path is the right one. One operator, the full open-source stack, built so you own what we ship.