WordPress hosting for agencies: ownership, margin, and handoff

First decide who owns the hosting

Decide whether the client will own and pay for hosting or whether your agency will provide it within a retainer. You can provide a professional service under either arrangement, but the billing, permissions, and eventual handoff differ.

Document who gets renewal notices, who can contact support, and who authorizes a restore. Do not share an owner login with every contractor. Use roles appropriate to the work and remove access when someone leaves.

What each platform contributes

WP Engine: evaluate its current site-transfer process, environments, and permissions. These can suit an agency that builds sites and hands them to client-owned accounts. Confirm any separately sold agency or billing products; a generic hosting subscription is not evidence of a white-label invoice portal.

Cloudways Flexible: evaluate how you would divide clients across servers and how those servers would be maintained. Cloudways has team and transfer capabilities; it is inaccurate to describe it as having no agency tools. A whole-server transfer is useful for a single-client server but does not separate one client from a shared portfolio.

Kinsta: evaluate the MyKinsta workflow with the people who manage releases and support. Judge it by the tasks you can repeat reliably across accounts and sites. A higher invoice may be acceptable if it reduces work you otherwise pay for, but the dashboard alone does not prove that saving.

Calculate margin per paying client

Start with revenue from hosting and maintenance, then subtract hosting, paid services, licenses, and labor. Do not treat the full retainer as hosting margin when it also includes content edits or support.

An illustrative ten-client portfolio charges $60 per client monthly, producing $600. Hosting and tools cost $180; four hours of maintenance at $50 cost $200. That leaves $220 before other business costs. If maintenance takes eight hours, it leaves only $20. The provider has not changed; the workload has.

The same principle applies to plan slots. A plan with twenty slots and ten paying clients should be divided by ten for current margin, not twenty. For rates and comparison arithmetic, use the pricing guide.

Three agency decisions, worked through

These are fictional budgets to demonstrate the arithmetic, not provider quotes, customer results, or performance tests. Each example compares the same sites for twelve months at steady renewal rates. The worksheet leaves service requirements unconfirmed: a lower number cannot establish that a plan can run your sites.

Five brochure sites: try a smaller current plan first

Suppose hosting costs $100 a month, extras $20, and your work takes two hours at $50 an hour. A smaller current plan costs $80 with the same extras and work, plus one hour to make the change. A different host quotes $60, but extras rise to $25 and your work to three hours; moving takes four hours and $100 in paid help. There is no prepaid credit to recover.

Year-one costs including your time are $2,640 staying, $2,450 changing the current plan, and $3,120 switching. The smaller plan saves $20 monthly and recovers its $50 setup effort in 2.5 months. The cheapest hosting quote creates the highest overall cost in this example.

Check that the smaller plan still covers storage, traffic, backups, and every required site. Do not downgrade because sites look simple; use their actual requirements. Open and edit the five-site comparison.

Six sites including a store: test the work before pricing the saving

Assume the current portfolio costs $180 monthly for hosting, $40 for extras, and four hours of your work at $60 an hour. There is $180 of usable prepaid hosting. A larger current plan costs $220 and takes two hours to change; it keeps that credit. For this calculation only, suppose it reduces monthly work to three hours. A new host costs $130 plus $60 of extras, six hours of monthly work, $180 in fees and five hours to move, with no credit recovered.

Year-one effective costs are $5,340 staying, $5,220 changing the current plan, and $7,080 switching. The larger plan saves $20 a month after valuing time and recovers the $120 of setup work in six months. If it fails to remove that hour, it instead costs $40 more each month.

The reduced workload is an assumption to test during a pilot, not a benefit promised by any host. Check store jobs, checkout, restore procedures, and whether the store needs a separate account or resources. Six sites alone says nothing about capacity. Open and edit the store-portfolio comparison.

A client handoff: ownership may matter more than the monthly bill

Here the budget covers one client's hosting and the work needed to administer it, whoever pays the invoice. Current hosting is $40 a month, extras $10, and work half an hour at $50 an hour, with $120 of prepaid hosting. Keeping the provider while changing the account arrangement has the same ongoing costs, $50 in fees and one hour of handoff work; it preserves the credit. Switching lowers hosting to $30, but costs $100 in fees and three hours of work and loses the credit.

Year-one effective costs are $780 staying, $880 changing the current arrangement, and $1,030 switching. Switching saves $10 monthly but takes 37 months to recover fees, work and lost credit. That assumes the same rates and work continue for more than three years.

A client who needs independent ownership may still choose the $100 handoff. Confirm the provider's transfer rules, access, billing and credit treatment before using that option. This compares costs, not agency profit: lost retainer revenue and the client's preferred ownership arrangement need a separate decision. Open and edit the handoff comparison.

Group sites by workload and consequence

Site count is an inventory measure, not a capacity estimate. Record which clients run stores, scheduled imports, memberships, large media libraries, or frequent deployments. Consider separating a demanding application from lightly used brochure sites.

Separation is not automatically redundancy. Two servers can still depend on the same account, region, or external service. Write down what happens if an account is inaccessible or a server fails, and keep the recovery instructions somewhere the team can still reach.

Pilot the whole client lifecycle

Before moving the portfolio, take one representative site through these steps:

1. Onboard the site, its domains, integrations, and required access.
2. Create staging and make a small approved release.
3. Check forms, customer actions, scheduled jobs, and email.
4. Restore a backup into a safe environment and verify the result.
5. Rehearse a client departure, including files, database, DNS, credentials, and billing.

Record actual time and anything requiring provider support. The developer guide covers release requirements; the migration guide covers the live cutover.

Which arrangement should you choose?

Choose WP Engine if its handoff and release process fits your team and the appropriate subscription makes financial sense. Choose Flexible when you have technical ownership of the server arrangement and useful portfolio economics after labor. Choose Kinsta when its managed workflow is the one your team can operate confidently at the quoted cost.

A split portfolio can be sensible: keep a demanding store separate while standardizing simpler clients elsewhere. Only add another platform when the benefit exceeds the burden of maintaining two procedures. Once you have tested the process, review the matching provider offer rather than moving every client based on a headline cost per site.