A founder sends us a screenshot of the Payload website and a one-line email: "It says free. Why is your quote €48,000?" Fair question. Payload CMS is MIT-licensed open source — you can clone it, run it, and never pay the project a cent. What you cannot do for free is design the content model, migrate 4,000 posts out of WordPress, build the Next.js frontend, host it under real traffic, back it up, and teach twelve editors to use it. That is the actual purchase. This article is the year-one budget we put in front of founders and CFOs before anyone signs anything — the same numbers, in the same shape, whether the prospect ends up hiring us or not. If you are budgeting a move off WordPress, Webflow, or Contentful onto Payload + Next.js, this is the spreadsheet your CFO will ask for in the second meeting.
The $0 license: what free actually covers
Payload's MIT license covers the framework: collections, fields, hooks, access control, the admin UI, the Local API, migrations. There is no per-seat fee, no content-tier paywall, no enterprise phone call required to unlock versioning. Compare that to Contentful, where a team plan with the locales and environments a serious editorial operation needs runs into five figures annually, or Sanity, where per-seat pricing compounds as the editorial team grows. This is the real economic argument for Payload, and it is genuine.
What is not free is everything around the framework. Payload offers Payload Cloud as a managed hosting option — check the current tiers on the official Payload pricing page before budgeting, because the plans have shifted over the past year — but most of the teams we work with self-host on Vercel + a managed Postgres, which keeps the infrastructure portable and the bill itemized. Either way, the license line on your year-one budget is €0. Every other line is real money.
The build is the budget: €25k to €80k, and what moves it
Across the Payload + Next.js builds we have scoped, year-one build cost lands between €25,000 and €80,000, and the spread is not arbitrary. Four scope drivers decide where you land:
Collection complexity — A marketing site with pages, posts, and authors is a week of modeling. A platform with 14 collections, localized fields, role-based access, and editorial workflows is four to six. Each collection with real relationships costs design time, not just config time.
Content migration volume — Moving 200 pages out of Webflow is a scripted afternoon. Moving 4,000 WordPress posts with shortcodes, embedded media, and broken slugs is 3–6 weeks of ETL, dry runs, and redirect mapping. Migration is consistently the most underestimated line.
Design ambition — A clean template-driven frontend on Tailwind is fast. A bespoke design system with animation, custom blocks per landing page, and art-directed templates can double the frontend budget.
Integration count — Every external system (CRM, newsletter, search, DAM, analytics, commerce) adds 3–10 days depending on the quality of its API. Six integrations is a different project than two.
In weeks: a marketing site replatform runs 4–8 weeks; a serious editorial platform with migration runs 8–14 weeks; a multi-tenant or docs-plus-marketing build runs 12–20 weeks. Any quote that promises the second shape in the first timeline is either padded with risk you will pay for later, or scoped by someone who has not done the migration.
Hosting at three traffic tiers
Self-hosted Payload + Next.js has a predictable monthly shape: Vercel for the app, Neon or Supabase for Postgres, Cloudflare R2 or S3 for media, Resend for transactional email. These are the ranges we see on real invoices, as estimates — your numbers will move with traffic and media volume, and the vendors update pricing often enough that you should verify against their current pages:
~50k page views/month (marketing site): Vercel Pro ~$20–$60, Neon launch tier ~$19, R2 near-zero egress ~$5–$15, Resend free–$20. Total: €150–€250/month with monitoring and a preview environment.
~300k page views/month (active editorial platform): Vercel with higher function usage ~$100–$250, Postgres scaled tier ~$50–$100, storage + bandwidth ~$30–$80, email ~$20–$50. Total: €350–€550/month.
~1M page views/month (high-traffic platform): Vercel usage-based ~$300–$500, Postgres with read replicas ~$150–$250, media + CDN ~$80–$150, email and extras ~$50–$100. Total: €700–€900/month.
Two honest caveats. First, Vercel is the right default for teams without DevOps capacity, but at sustained high traffic a container on Hetzner, Fly.io, or AWS ECS can cut the compute line meaningfully — we have seen 40–60% reductions at the top tier, at the cost of owning deployments and failover. Second, Payload Cloud bundles some of this; run both totals before deciding, because convenience and portability pull in opposite directions.
The line items teams forget
Every under-budgeted Payload project we have been called in to rescue forgot the same four things. None of them are expensive. All of them are painful when missing.
Preview environments — Every PR should spin up a preview so editors can review content changes against the real frontend. On Vercel this is nearly free; forgetting to scope the wiring is a week of retrofitted work.
Backups with actual restore tests — A nightly Postgres dump nobody has restored from is a hope, not a backup. Budget the quarterly restore drill: ~2 hours of engineering time, or the first real incident teaches you the difference at the worst moment.
Monitoring and error tracking — Sentry or equivalent, uptime checks, and log retention: €30–€80/month. The cost of not having it is learning about outages from your customers.
Editor training — Half a day with the editorial team, plus a short internal playbook. This line is under €1,000 and it decides adoption. We have watched a €60k platform sit unused for two months because nobody showed the editors how preview drafts work.
After launch: retainer, hire, or ad-hoc
Year two is where the total-cost comparison with SaaS CMSs gets interesting, because Payload has no license fee but does have a maintenance reality. Three options, honestly priced:
Studio retainer — €1,500–€4,000/month for a standing allocation: dependency upgrades, Payload version bumps, small features, incident response. Right for teams shipping changes weekly.
In-house hire — One mid-senior TypeScript developer at €60k–€85k fully loaded. Only makes sense past roughly 20–30 hours of platform work per month; below that you are paying for capacity you do not use.
Ad-hoc support — €100–€150/hour when something breaks. Fine for a stable marketing site; dangerous for an editorial platform, because the person who knows your codebase is never the one available at 9pm.
What we tried on similar builds: the pattern that holds up is a small retainer for the first six months post-launch — when the backlog of "while we're in there" requests is longest — then a deliberate step down to ad-hoc once the platform stabilizes and the editors are self-sufficient. Teams that skip the retainer phase almost always come back for it after the first Payload major-version upgrade lands.
When we talk prospects out of Payload
Roughly one in five Payload inquiries we receive ends with us recommending something else. The three recurring cases: the five-page marketing site that updates twice a year — a €29/month Webflow or Framer plan is the correct answer, and a headless CMS is vanity infrastructure. The team with no developer access — Payload is a framework, not a product; if nobody in your orbit can run `pnpm dev`, the €0 license buys you nothing. And the simple blog that could live on Ghost or a static generator for a tenth of the total cost. We would rather lose the project than win a client who will resent the invoice in month three.
If the alternative you are weighing is a SaaS headless CMS rather than a site builder, we published the head-to-head numbers: Payload vs Sanity for content teams: the 3-month cost breakdown.
Three sample year-one budgets
These are the three shapes we quote most often, with every line item shown. Numbers are planning ranges from projects we have scoped, not a rate card — but they are the right ranges to sanity-check any quote against.
Scenario A — Marketing site (8–12 pages, blog, 2 integrations, light migration): build €25k–€35k, hosting €150–€250/month, training + extras ~€2k, year-two maintenance ad-hoc. Year one: €30k–€42k.
Scenario B — Editorial platform (40 posts/month, 8–10 collections, 3,000-post migration, 4 integrations): build €45k–€65k, hosting €350–€550/month, training + monitoring ~€4k, 6-month retainer €12k–€18k. Year one: €65k–€95k.
Scenario C — Multi-tenant marketing + docs (multi-tenant collections, versioning, docs search, 5+ integrations): build €70k–€80k+, hosting €500–€900/month, retainer €18k–€24k. Year one: €95k–€130k.
For comparison, the Contentful or Sanity license alone for Scenario B's team size typically runs €10k–€30k/year before you have built anything — which is why the three-year TCO comparison usually favors Payload even when year one does not. If your content operation also touches commerce — a catalog, a storefront, a checkout — the same self-hosted economics apply on that side of the stack too.
When the platform budget includes a commerce layer alongside the content platform, See how we ship Medusa storefronts — the build-scope logic above maps almost line for line.
How to read any agency quote for a Payload build
Six questions expose padding or under-scoping in any quote, ours included. Ask them verbatim:
How many collections and custom blocks are in scope, and what is the per-collection change cost after signing?
Is content migration a fixed line or T&M — and how many dry runs are included before cutover?
Which integrations are quoted, and what happens when a vendor API changes mid-build?
Are preview environments, backups, and restore tests itemized, or assumed?
What does the quote assume about design — bespoke system, adapted template, or our existing brand kit?
What is explicitly out of scope? (A quote without an out-of-scope section is a quote that will grow.)
What we would do for a client in this position: before signing anything, map your project against the three scenarios above, get the hosting tiers on one spreadsheet line each, and force every quote through the six questions. If you want a second set of eyes on a quote you are already holding — ours or anyone else's — that is a conversation we have for free, because it takes twenty minutes and usually saves five figures.
Send us your replatform brief — we will stress-test the budget