Agencies running 20+ client sites usually default to whatever hosting they started with and never revisit the decision. That's expensive. The right choice depends on how much control you need over each site, how much per-site cost you can absorb, and whether your sites are still separate WordPress installs or already consolidated onto shared infrastructure -- and the gap between the cheapest and most efficient option can run into tens of thousands of dollars a year.

Key takeaways

  • Cost per site drops fast once you move off one-account-per-site shared hosting -- the real savings come from infrastructure you share across sites, not from a cheaper per-site plan.
  • A VPS gives you full control and predictable pricing at 20+ sites, but you own patching, isolation, and scaling yourself. Multi-region providers matter if your clients' traffic is geographically spread.
  • Managed WordPress agency plans (Pantheon, Cloudways, Liquid Web, WP Engine and similar) trade some control for a single dashboard and someone else owning uptime -- worth it below roughly 20-30 sites, less so above it.
  • Headless/Jamstack hosting (Vercel, Netlify, Supabase) is the cheapest option per site once you're past a handful of client projects, because compute, database, and CDN are shared across every site on the account rather than duplicated per install.
  • The architecture decision matters more than the hosting decision -- 20 separate WordPress installs on great hosting still cost more to run than 20 sites on a shared multi-tenant backend.

Updated 25 August 2026: published.

What "Hosting for 20+ Sites" Actually Means

At 20+ sites, hosting stops being about picking a plan and becomes an operations decision. You're not asking "which host is fastest" -- you're asking who patches 20+ cores, who gets paged when one client site gets hammered by a traffic spike, and whether a security issue on Client A's install can touch Client B's data.

Three variables decide the right answer for your agency:

Isolation requirements. Some client contracts (healthcare, finance, anything under a compliance regime) require hard isolation between tenants. Others don't care, and shared infrastructure is fine.

Team size relative to site count. A five-person agency running 20 WordPress sites needs different tooling than a fifteen-person agency running the same 20 sites, because the second team can absorb more manual ops work per site.

Whether the sites are still separate installs. This is the variable that actually determines cost, and it's the one most agencies never revisit once they've onboarded past client #5.

The Real Cost Comparison

Here's what changes as you move through the options, based on real migration numbers rather than list prices.

Shared or reseller hosting, one account per site. This is where most agencies start and where the cost problem compounds. Every site gets its own control panel, its own database, its own SSL renewal, its own plugin stack drifting slowly out of sync with every other site. We wrote up the actual math on this at 50 WordPress sites: $36,000 to $78,000 a year in hosting, maintenance, and security overhead once you count the hours spent keeping 50 separate installs patched and working. That number scales down roughly linearly, not to zero, at 20 sites -- you're still running 20 separate cores.

A VPS, self-managed or lightly managed. One server (or a handful, split by client tier) hosting multiple sites gives you predictable monthly pricing that doesn't scale per-site the way shared hosting does. You take on patching and isolation yourself, or pay a provider for a managed tier. Where your server physically sits matters here too: multi-region VPS providers like DigitalCloud.pro let you place client sites closer to their actual audience instead of hosting everyone from one region and eating the latency, which matters more than most agencies expect once Core Web Vitals become a client SLA line item.

Managed WordPress agency plans. Pantheon, Cloudways, Liquid Web, WP Engine and similar products exist specifically to solve the "20 WordPress installs" problem without a rebuild -- one dashboard, staging environments, and someone else on call for uptime. That's real value, and for agencies under roughly 20-30 sites who aren't ready to change platforms, it's often the right call. The tradeoff shows up as you scale past that: pricing is still fundamentally per-site, so the cost curve doesn't bend the way it does once you're off separate WordPress cores entirely.

Headless or Jamstack hosting. This is what we actually migrate agencies and their clients onto, so the numbers here are from our own production stack, not a vendor's pricing page. A single project's hosting on Supabase + Vercel runs about $540/year -- against $10,000 to $30,000/year for the equivalent WordPress Multisite network hosting, or $250,000 to $500,000/year at the Sitecore/AEM enterprise end. The bigger shift at 20+ sites is architectural: a multi-tenant Next.js + Supabase setup serves many client sites from one shared backend, so compute, database connections, and CDN caching aren't duplicated per site the way they are on separate WordPress installs. We covered the multi-tenant vs. multi-site decision in more depth here, including the routing pattern in our wildcard-subdomain Next.js middleware writeup.

Two real examples from client work: SleepDr.com went from a WordPress install scoring 35 on Lighthouse mobile to 94 after migrating to Next.js 15 with Payload CMS and Supabase. bdManagedIT, a Central-Georgia managed IT provider, moved off WordPress onto Astro and Sanity with Netlify hosting and landed on 95+ PageSpeed with zero-JS static pages across its service, industry, and location pages. Neither number is a hosting change alone -- it's what happens when hosting and architecture get fixed together.

When a VPS Is the Right Call

A VPS makes sense when you want infrastructure ownership without a full platform rebuild, and your sites are still going to be separate installs for the foreseeable future -- whether that's WordPress, a mix of stacks across clients, or anything that doesn't fit neatly into one shared backend. It's also the right call when specific client sites have unusual resource needs (a high-traffic ecommerce client sitting next to five low-traffic brochure sites) that a flat per-site managed plan prices inefficiently.

Multi-region matters more than agencies think it does. If your client roster spans continents -- a UK client, a US client, an APAC client -- a single-region VPS puts two of those three sites at a real latency disadvantage on every request, which shows up directly in TTFB and, downstream, in Core Web Vitals.

When Managed WordPress Agency Plans Make Sense

If you're under roughly 20-30 sites, all on WordPress, and your team doesn't have spare engineering hours to own a VPS or a migration project, a managed agency plan is a legitimate answer, not a compromise. You're paying for someone else's on-call rotation and staging tooling, and at that scale the per-site premium is smaller than the cost of your own team's time.

The math changes as the site count climbs and as any individual client outgrows what a shared managed-WordPress tier can comfortably serve -- that's usually the first signal worth investigating a platform change for that one site, even if the rest of the portfolio stays put.

When It's Time to Go Headless

The honest trigger isn't a site count -- it's whether you're already paying the WordPress-at-scale tax described above and have the appetite for a migration. Agencies that move tend to do it client-by-client rather than all at once: migrate the highest-traffic or highest-maintenance site first, prove the hosting-cost and performance delta on a real project, then use that case to justify moving the rest of the portfolio.

A Simple Decision Framework

  • Under 10 sites, mostly WordPress, no compliance pressure: shared or a light managed plan is fine. The overhead isn't big enough yet to justify a change.
  • 10-30 sites, still WordPress, team has no spare engineering time: managed WordPress agency hosting. Pay for the dashboard.
  • 20+ sites, or a handful of sites eating disproportionate engineering time, and you're comfortable owning infrastructure: a VPS, ideally multi-region if your clients aren't all in one geography.
  • 20+ sites and the WordPress-at-scale costs are visibly eating margin: start migrating the worst offenders to a shared multi-tenant backend. This is the option with the largest long-term cost delta, and the one most agencies underrate because it looks like the most work up front.

FAQ

What's the cheapest hosting for an agency with 20+ client sites?

Per-site cost is usually lowest on a shared multi-tenant backend (one Next.js/Supabase project serving many client sites), because compute and database aren't duplicated per site. A VPS is the cheapest option that keeps sites as fully separate installs, since pricing doesn't scale per-site the way shared or managed hosting does.

Is a VPS better than managed WordPress hosting for agencies?

It depends on your team's spare engineering time. A VPS is cheaper at scale and gives you full control, but you own patching, backups, and isolation yourself. Managed WordPress hosting costs more per site but includes that ops work, which is worth it for smaller teams under roughly 20-30 sites.

How much does it cost to host 20 WordPress sites?

There's no single number -- it depends on hosting model. On separate shared installs, agencies typically see $36,000 to $78,000 a year in combined hosting, maintenance, and security overhead once 50 sites are involved; that scales down but doesn't disappear at 20. Managed WordPress plans and VPS options both reduce that, and a shared multi-tenant backend reduces it the most.

Does server location actually matter for client site performance?

Yes, more than most agencies assume. If your VPS or hosting region doesn't match where a client's traffic actually comes from, that shows up as added latency on every request, which affects Time to First Byte and downstream Core Web Vitals scores -- a real factor if a client's contract includes a performance SLA.

Should every client site move to the same hosting stack?

Not necessarily. Compliance-sensitive clients may need hard isolation regardless of what's most efficient for the rest of your portfolio. The framework here is about the majority of your book, not a rule that every single site must match.

What's the difference between multi-tenant and multi-site hosting for an agency?

Multi-site means each client gets a separate install (even if they share a host). Multi-tenant means multiple client sites share one backend -- one database, one deployment, isolated by tenant ID rather than by separate infrastructure. Multi-tenant is cheaper to run at scale but requires a platform built for it; retrofitting WordPress Multisite for this is possible but rarely clean.

When should an agency consider migrating off WordPress entirely instead of just switching hosts?

When the $36K-$78K-a-year math from managing separate WordPress installs is bigger than the cost and disruption of a migration, and when it's recurring rather than a one-time pain point. Most agencies find the answer clearer after migrating one client site and comparing the real before-and-after hosting bill.