
A WordPress maintenance business sells a dependable operating process: know the site, make recoverable changes, verify the important functions and explain the result. Clicking an update button is only a small part of that work. The opportunity is useful technical support for an owner who cannot or does not want to manage the process personally.
Start with sites and tasks you can support competently. A small publication is a different responsibility from a busy store, a membership service or a site handling sensitive information. If you have only maintained your own blog, do not immediately sell unrestricted emergency support for every type of WordPress installation.
This guide explains how to define a first maintenance offer, test the workflow and price the responsibility. It is not a guarantee that a site will never fail or be compromised, and it is not a substitute for specialist security or development work when those are needed.
Inventory the site before promising a package
A discovery checklist should establish the host, domain and DNS ownership, WordPress setup, themes, plugins, custom code, backup arrangements and critical integrations. Ask which parts of the site generate business and what happens if they stop working.
An apparently simple brochure site might rely on a contact form, a booking service and an external email account. Its owner may care more about one working inquiry form than a perfect-looking homepage. Identify those dependencies before designing the maintenance checklist.
Ask who can authorize changes and who holds the necessary accounts. Do not begin by collecting every password in a shared document. Use supported user invitations and scoped access where possible, and agree on how access will be removed when the relationship ends.
Record unsupported components and known problems separately. Your first invoice should not silently turn a pre-existing, undocumented issue into a promise that you have fixed it. If the site needs a repair or migration before routine maintenance can begin, scope that as its own job.
Define routine maintenance and exclusions
A useful package states what you check, how often you check it, what updates you handle and what reporting the owner receives. It also explains support hours, response targets and the boundary between maintenance and development.
For example, a basic publication package might include planned update review, backup checks, an agreed functional checklist and a concise monthly report. New page design, content entry, malware cleanup, major compatibility repairs and round-the-clock support might be excluded or separately scoped.
Be precise about response time. A promise to acknowledge a report within a business day is not a promise to resolve every problem within that period. Specify what qualifies as an incident, how it should be reported and what happens outside your available hours.
Avoid offering “unlimited fixes” unless you have a credible operating and pricing model for them. The phrase sounds reassuring but can conceal an agreement neither side understands. A defined scope with a sensible escalation route is more useful to a small client.
Build a recovery process before an update process
WordPress’s backup documentation explains that a typical full-site recovery requires both files and the database. A copy of uploaded images alone will not restore the site’s content and settings, and a database export alone will not replace all necessary files.
For each client, establish what is backed up, where copies live, who controls them and how long they are retained. Consider what happens if the hosting account becomes unavailable. A backup stored only inside the account it protects may be inaccessible during the incident that makes it necessary.
Set recovery expectations with the owner. How much recent data could be missing from a restored copy? How long might restoration take? A weekly content snapshot may be appropriate for one small site and inadequate for a site taking orders throughout the day.
Test recovery in an isolated environment that will not send real emails, process live payments or confuse visitors. Keep access and indexing controlled. A backup job’s success message is useful evidence that something ran; a successful restore is stronger evidence that the required data is usable.
Document the test and its limits. If the backup excludes an external booking service or a third-party mailbox, say so. Do not describe the website backup as a complete business recovery plan when it covers only one part of the operation.
Rehearse a change on a suitable copy
The official WordPress upgrade guidance emphasizes backups before an upgrade. Your service should also consider compatibility, the site owner’s change window and a way to verify the functions that matter after the change.
When feasible, use a controlled staging copy to assess higher-risk changes. Protect personal information and block unwanted live integrations. Staging is not automatically safe merely because its address contains the word staging; check what emails, webhooks, scheduled jobs and external services it can reach.
Create a short change plan. Record the current versions, the proposed change, the expected effect and a rollback decision point. Grouping many unrelated changes into one unexplained action can make a later failure difficult to diagnose.
Some updates are urgent; others need more preparation. Define how you will evaluate urgency within your competence and escalate specialist questions. Indefinitely postponing all changes is not a maintenance strategy, but applying every change immediately without understanding the site is not a reliable one either.
Check business functions rather than only the homepage
Build the functional checklist with the owner. A publication might require article pages, search, images and editor login. A service business might require forms, booking links and delivery to an appropriate test destination. A store adds responsibilities that demand suitable skills and controlled test methods.
| Check | What it can reveal | Important limitation |
|---|---|---|
| Public page request | A page is reachable and renders | Does not prove every feature works |
| Login and editor access | An authorized user can reach the workflow | Does not prove all permissions are correct |
| Approved form test | A particular submission path completes | Does not prove every inbox accepts delivery |
| Backup verification | Expected data is present or restorable | Covers only the tested recovery scope |
| Mobile layout review | Visible problems at tested screen sizes | Does not establish complete accessibility |
Use authorized test accounts and agreed test data. Do not submit a real order, contact a real customer or send a message to an unrelated person just to prove a feature works. A maintenance contract should identify suitable test destinations and who approves actions with external effects.
If a test cannot be performed safely, record that limitation. “Not tested because no approved recipient was available” is more honest and useful than a green check mark based on assumptions.
Understand why rollback can become a data problem
Restoring an old database after the live site has accepted new orders, comments or member changes can discard legitimate activity. That is why a rollback plan needs more thought than “restore yesterday’s backup if anything looks wrong.”
For a planned change, decide whether a brief write freeze is needed and how you will manage it. If the site cannot stop accepting changes, consider whether you have the expertise and tooling to preserve or reconcile new data. Escalate rather than improvising on a business-critical installation.
Keep the previous software or configuration available when your deployment method supports that. Record what can be reverted independently and what requires data restoration. An application image rollback and a database rollback are related but different operations.
If failure occurs, preserve relevant logs and avoid repeated uncontrolled changes. Communicate what is known, which functions are affected and what you are doing next. Do not promise a recovery time based only on optimism.
Recommend tools according to the operating need
The WordPress security handbook provides a reference for maintaining the installation and limiting unnecessary exposure. A security plugin or managed host can support a process, but neither removes the need for appropriate access, updates, recovery and competent decisions.
When comparing hosting, consider backup access, restore controls, support boundaries, staging options, account permissions and the renewal price of the required plan. SiteGround’s WordPress hosting information is one product page to evaluate alongside the client’s existing host and other suitable providers. This is not a claim that it is the best host or that a particular plan has every feature you need.
Prefer the existing setup when it adequately supports the agreed service. A migration introduces its own risks and costs. Do not recommend one merely because a provider offers a referral payment or an introductory discount looks attractive.
For backup, monitoring or management tools, evaluate what their failure would mean. Can you export the data? Does the owner retain access? Does one account control all clients? What happens if you stop paying? Those questions are more useful than counting features on a comparison chart.
Price the work and the availability separately
Estimate routine administration, planned changes, checks, reporting and a sensible allowance for support within the defined package. Include software costs, client onboarding and the time needed to remain competent. Availability itself can limit your ability to take other work.
In a hypothetical package, a $120 monthly fee with $20 of allocated direct tool costs leaves $100 before other overhead and taxes. If the package uses two hours in an ordinary month, the result is $50 per hour on that narrow calculation. One additional five-hour incident changes it to roughly $14.29 per hour across seven hours unless that incident is separately scoped.
This is why exclusions and incident pricing matter. It is not a recommendation to charge $120, nor evidence that clients will accept a particular price. Design a package whose obligations you can actually meet, then test demand with suitable prospects.
Review unusual sites individually. Multiple languages, custom checkout, membership access and unsupported software can change both the workload and consequences of error. A single flat price for every WordPress site may hide those differences until a difficult month exposes them.
Make the monthly report explain decisions
A useful report names the work completed, the important checks performed, known limitations and decisions needed from the owner. Avoid burying the client under a raw plugin version list or presenting every blocked bot request as a prevented catastrophe.
Distinguish observation from diagnosis. “The booking page returned an error during the scheduled check” is an observation. “The booking provider is responsible” requires further evidence. Clearly state when an issue remains under investigation.
Keep a change log detailed enough for another competent person to understand the sequence. Record recovery copies, account ownership and essential handoff information in the agreed secure location. Your service should improve the owner’s control over the site, not make basic access dependent on your memory.
Practice an incident decision before a real incident
Consider a fictional appointment-based business whose site stops confirming bookings after a planned update. The homepage still loads, and the owner initially assumes the whole maintenance check passed. Your agreed functional checklist reveals the failure before you describe the work as complete.
First, establish what is actually failing. Does the form reject the request, does the booking fail to appear in the external system, or does only the email confirmation go missing? Those observations lead to different investigations. Changing DNS, restoring the database and reinstalling a plugin all at once would make it harder to identify the cause.
Check the change record and the relevant logs within your access and expertise. Preserve the current evidence. If the problem appears related to a recent component change, assess whether a software rollback can be performed without discarding new booking data. If not, involve the appropriate developer or provider before taking a destructive shortcut.
Meanwhile, communicate the limitation accurately. Explain which booking path is affected and ask the owner to approve any temporary alternative shown to visitors. Do not silently redirect customer information to an account you control because it is convenient for testing.
After service is restored, repeat the approved end-to-end check and explain what remains unverified. Record the resolution and whether the maintenance checklist or change window should improve. An incident review should produce an operational change, not merely a reassuring paragraph in the next report.
You can rehearse this reasoning on a demonstration site without creating a real business disruption. Practice recognizing when the problem exceeds your competence and how you would hand the evidence to someone qualified. Knowing when to escalate is part of a useful technical service.
Keep onboarding and exit checklists equally usable
At onboarding, establish account ownership, contact details, escalation channels and a current inventory. At exit, provide the agreed records, remove access appropriately and explain where backups and licenses reside. Both checklists should be understandable to another competent person.
Clarify what happens to paid tools supplied under your business account. A client should know whether a license continues, must be replaced or stops receiving updates when the relationship ends. Avoid presenting a low monthly package that quietly leaves the site dependent on your account without a handoff plan.
For external monitoring, identify which alerts go to you and which go to the owner. A monitoring dashboard that nobody checks outside your working hours does not provide round-the-clock response. The report should distinguish automated observation from human availability.
Review the agreement periodically as the site changes. A publication that adds paid memberships or a store has changed its operating needs. Update scope and recovery expectations before treating the new features as if they were always part of the original package.
Win a first client through a limited engagement
Begin with a paid assessment or a narrowly defined maintenance pilot for a site you can support. Show a sample report using fictional information and explain the boundaries. Do not scan or modify someone else’s site without authorization to create a sales opportunity.
If the owner wants a redesign, ecommerce repair and migration alongside routine support, separate those projects. It is easier to build a healthy recurring relationship when both parties understand what the initial work establishes.
For guidance on defining a small service and approaching appropriate prospects, see our first-client guide for virtual assistants. The technical responsibilities differ, but the principle of selling a specific, reviewable result is transferable.
Evaluate the pilot by completed work, client communication, actual time and recovery readiness. A small number of well-understood sites can be a more manageable business than a large collection of inexpensive accounts with unclear obligations. Expand only when your process and capacity support the added responsibility.
Research checked November 16, 2026 UTC. Product pages describe options to evaluate, not personally tested rankings. All pricing examples are hypothetical; no service can guarantee uninterrupted availability or complete security.
Cover illustration generated with AI.
This guide is educational and uses AI-assisted research and drafting. Examples labeled illustrative are calculations, not earnings reports. Check current source terms before acting. Read our editorial approach.