Moving a WordPress Site to a New Host: The Ownership and Handover Checklist

By Slut Teens 2026-10-26

Brand promotion: this article introduces Guangsuan's WordPress hosting and speed optimization service. It is a vendor-introduction piece, not an independent review, and it was not commissioned by a customer.

Moving a WordPress site to a new host is a two-part job, and the parts are usually confused. The first part is technical: copying files and a database across, pointing the domain at the new server, checking that pages and forms still work. The second part is ownership: deciding who holds the domain, the DNS zone, the business mailbox, the analytics and search accounts, and the backup copies — and making sure each of those can still be reached when the old provider stops answering.

This checklist covers the second part, with the technical steps included only where they touch it. It is written for a site owner or operator who already runs a WordPress site and is deciding between moving to a different host and handing the site to someone else to look after.

The short answer

Before you switch, get four things settled in writing: which accounts the business holds, which records the cutover touches, what a copy of your data actually contains, and what you will measure the new setup with. Everything else in the migration follows from those four answers.

Two conditions belong with that answer. First, this applies to a site that already exists and already has its own domain and a business email address on it; a site that has never been live has a different conversation. Second, none of the four items transfers automatically. A site can move cleanly, open correctly on the new host, and still be held by accounts the business does not own.

1. The domain and the DNS zone: two different accounts

Domain registration and DNS hosting are separate, and a handover that covers one can leave the other behind.

Ask who holds the registrar account — the login that can renew, unlock or transfer the domain — and who holds the DNS zone, which is where the actual records live. They are often the same company, and they are often the same company as the old host. Confirm the renewal method: which card or balance pays, and which email address receives the expiry notice. A domain that renews automatically from a former contractor's card is a domain the business does not fully control.

Before you change anything, write down the records that matter and what each one is for:

The A or AAAA record for the domain, and the www CNAME, which point at the host serving the site.
The MX records, which route incoming email to your mail provider. If your mail host is a subdomain such as mail.example.com, check that its address record is right too.
Sender-authentication records published by your mail provider — SPF, DKIM and DMARC — and any domain-verification record for a service you use.
Any verification record for a search console or analytics property.

The key question to ask the current provider is whether you are changing only the website records, or replacing the nameservers that host the whole zone. Changing the website records is a smaller job. Replacing the nameservers means every record has to be recreated in the new zone before the switch, and a record you forget becomes a problem you diagnose later, under time pressure.

One caution: do not cancel the old domain, host or mail service until the new arrangement has been verified. Changing records and waiting is reversible. Letting a registration lapse, or closing a mailbox while messages are still arriving, is not.

2. Business email: three layers, and only one of them is the inbox

An email address printed on your contact page is part of the business identity, and it has three layers that are easy to mix up: the mailbox itself, the administrator console that manages the mailboxes, and the recovery route for that console.

At handover, confirm:

Which account owns the mail service and who is billed for it.
Who can create, suspend and delete mailboxes, and whether that console belongs to the business.
Which recovery email address and second factor protect the administrator account — and whether the business controls both.
Which aliases, distribution lists or forwarding rules exist, and whether any forward mail to an address outside the business.

That last item is worth checking at the same time as the contact form. A forwarding rule that sends every message to a former team member's personal address can keep working long after the site has been handed over, and the site looks fine the whole time.

Then test delivery rather than assuming it. Send a message to the address on the contact page, confirm it arrives in the mailbox the site names, and reply from that mailbox. A form that reports success and a mailbox someone can reply from are two different things.

3. Analytics and search accounts: reporting access is not ownership

Measurement accounts are business records even though they hold no pages. Two access questions matter more than the rest.

For analytics, the business should hold an administrator-level role on the property that collects the data, and should know which account owns it. If the property sits inside a contractor's account, you may be able to view reports and still be unable to keep the history or grant access to someone else later.

For a search console property, check how the property was verified. Verification by a DNS record proves control at the domain level and survives staff changes better than a verification that depends on one person's access to a file on the site. Either way, at least one owner should sit with the business, and the property should cover the address that is actually live.

Ask for a data export as well. Access to a report is not a copy of the data, and changing platform later is much easier with a dated export in hand.

4. Backups: a copy is not a recovery plan

A WordPress backup usually has two halves, and a copy with only one of them is not a usable copy: the database, and the uploaded files. Media, themes, plugins and site configuration live on the file side. A copy of only the website directory will not contain the database; a copy of only the database will not bring back the images.

Four questions cover most of it:

What is covered. Files and database, or one of them?
How often, and for how long. A mistake or an intrusion is not always noticed the same day, so the most recent copy may not be the useful one.
Where the copies live. A different datacenter is a different kind of protection than another folder on the same server. If several copies can be deleted through one account, the risk is not actually spread.
When a restore was last rehearsed. And in an isolated environment, rather than over live data.

Guangsuan's published hosting scope describes multi-datacenter offsite backups for exactly this reason, stating that site files and databases are not kept only on one server in one datacenter, so that an offsite copy reduces single-point failure risk and keeps a basis for recovery. The same page is explicit that the backup frequency, retention period, recovery method and recovery target are confirmed before activation, and that having backup files is not the same as having completed a restore.

5. Cache and CDN: agree the boundary before the switch

The cache layer is the part of a move most likely to produce a confusing first impression. A site can look fast on one request and slow on the next, because the two requests took different paths through the cache and the origin.

Three layers move at different speeds, and it helps to name them before you sign rather than after:

Server resources. The shared-machine specification, and how your site's share of it is defined. Guangsuan publishes that its usual host machines run at 16 cores and 32GB of memory or more with bandwidth above 100Mbps, and presents that as headroom for daily operation and traffic peaks rather than one dedicated machine per site. The same page labels those numbers as a host configuration reference rather than a site speed test, and notes that per-site allocation of resources, bandwidth and traffic follows the signed plan.
CDN. A CDN distributes static resources such as images, CSS and JavaScript across nodes in different regions. Guangsuan states that global CDN setup and configuration are included within the scope agreed in the plan, while node count, coverage, traffic and protection policy depend on the tier.
Page cache and bypass rules. A page cache stores the generated response for a public page so the server does not rebuild it on every visit. That is close to free speed on a brochure site and a hazard on a store, where the page a visitor sees depends on who they are and what is in their basket. Guangsuan's published scope includes WordPress cache and plugin optimization in all three tiers, with logged-in sessions, carts and checkout requests handled separately rather than cached.

The question to ask is therefore not whether there is caching, but which requests are cached, which are deliberately bypassed, and what happens to the cache when you publish a change. Ask to see the bypass list. On a store that usually means cart, checkout, payment callbacks and logged-in sessions; on an inquiry site it means forms and the pages that read them. Then ask how a replaced image or an edited price is purged.

6. Re-test under the same conditions, or the numbers mean nothing

A speed figure is only comparable with another figure produced the same way, so agree the measurement conditions before the work starts rather than after the numbers look good.

The useful minimum:

The same page, region, device, network limits and tool version for the baseline and the re-test.
Login state and cache state recorded.
Repeated runs kept individually before anything is summarised, so you can report a median and a spread rather than a single best number.
Laboratory results and real-user data presented separately, and not mixed into one figure.
Cold-cache, warm-cache and cache-bypass requests recorded separately on a cached site, because the same URL can look fast or slow depending on which one you measured.

Two metrics get confused in this conversation. TTFB is the wait from request to first byte, with the tool's timing definition, redirects and cache layer noted, and it is not the same as PHP execution time. INP needs real-user data or an actual interaction measurement; with no data it is marked as not yet measured rather than estimated.

Guangsuan's published approach is a formal optimization report tied to the test conditions, the basis of the diagnosis, the description of the change and the re-test record. The same page states that it does not promise a perfect score on every page, because theme weight, plugins, images, advertising, analytics scripts and the location of the test all move the number.

What to bring to the first conversation

Most of these conversations go badly for a simple reason: the provider has no idea what your site is, and you have no idea what to ask. Guangsuan's own guidance for an initial discussion is the site address, the main customer regions, the current host situation and the symptom you are seeing, followed by a free health check before the scope is confirmed.

Add the items below and you can hand the same sheet to any provider you are comparing:

Your pages. Two or three URLs you actually care about, plus one page that is unusually heavy.
Theme and plugin list. Usually where page weight lives, and it changes what a migration involves.
The paths that must not break. For a store: cart, checkout, payment callbacks and order emails. For an inquiry site: the forms and the inbox they reach.
Where your customers are. Region drives server placement and CDN decisions, so it belongs in the first message rather than the second.
Your measurement conditions. Which page, region, device and tool you will both use at acceptance.
Access, by an agreed method. Guangsuan's guidance is that backend and server access should be provided through an agreed secure method, and passwords should not be sent in a public channel.

Three tests worth running before the final invoice

Each of these produces evidence rather than a promise:

Create a limited editor account, publish or edit one low-risk page with it, then revoke the account. This proves the business can add and remove access without depending on the builder.
Submit a test enquiry through the live site and confirm it arrives in the mailbox or system the site names, with the fields intact.
Restore a backup in an isolated environment, or watch the provider do it. A restore that has never been run leaves you with a copy, not a recovery plan.

Keep the results with dates and the name of the person who ran them. That record answers a question that arrives months after go-live: who do we ask when this breaks?

What Guangsuan provides

Guangsuan provides WordPress hosting and speed optimization for export-oriented sites and WooCommerce stores, covering the technical side from onboarding through daily maintenance. Its published scope groups seven services around security, performance and ongoing support, and states that all three annual tiers include them, with the specific resources confirmed by the plan chosen. In practice that scope includes multi-datacenter offsite backups and dedicated security operations for a site that is hacked, attacked or running abnormally; host machines typically at 16 cores and 32GB of memory or more with bandwidth above 100Mbps, with per-site allocation set by the signed plan; global CDN configuration with tier-dependent node count and coverage, plus WordPress cache and plugin optimization in all three tiers; free migration covering the current-site check, backup, pre-migration, testing and the switch; 7×24 Chinese-language technical support for a site that will not open, an attack, a plugin conflict or a configuration change; and a formal optimization report with test conditions, diagnosis basis, change description and re-test record.

Two boundaries belong with that description. Migration is published as free and provider-handled, but two questions stay open until you ask: whether your own data and licence dependencies are covered, and whether a site that takes orders continuously gets a final data-synchronisation window instead of an unannounced cutover. And hosting is not a ranking or growth guarantee. Performance and stability are the foundation of site experience and technical SEO, while rankings also depend on content, search intent and competition, and inquiries depend on the product, the traffic quality and the page.

One practical condition is worth confirming early rather than late: the published technical support for the hosting service is described as 7×24 Chinese-language support. If nobody on your side reads Chinese, confirm the working language of the incident channel before you sign.

The scope, the tiers and the re-testing standard are set out together on Guangsuan's WordPress hosting and speed optimization page, which also offers a free check of your current site before you confirm a plan.

Frequently asked questions

Do I need to own the domain, or is access enough?

Ownership is the safer position when the domain is a long-term business asset. Access can be withdrawn, and an account held by someone else can be closed, transferred or lost. If the domain must stay with a provider for contractual reasons, put the renewal responsibility, the recovery contact and the exit route in writing.

Can the old provider keep administrator access afterwards?

Yes, if the business has decided that and can revoke it. A workable arrangement is a business-held owner account plus a named administrator account for the maintainer, reviewed when the contract ends. The risk is not the builder having access; it is the builder being the only route to recovery.

What if the analytics or search console property sits in someone else's account?

Ask for owner-level access, or for the property to be transferred, before the project closes, and take an export of the historical data. Recreating a property later does not restore the earlier history.

Is a backup enough to prove the site can be recovered?

No. A backup is a copy; recovery is a procedure with a tested result. Ask what the copy includes, how often it runs, how long it is kept, who can request a restore, and when one was last rehearsed. Guangsuan states that backup frequency, retention, recovery method and recovery target are confirmed before activation, which is the form of answer worth asking any provider for.

Will moving hosts improve my rankings?

No provider can promise it. Guangsuan's published limit is explicit: the hosting service is not a ranking or growth guarantee. What you can reasonably expect is that the site stays online, stays recoverable, and is measured the same way before and after, so you can see what actually changed.

What should I check on the day the site goes live?

The live domain rather than the staging address, the certificate, the contact route with a real test message, the redirects from any old addresses, and the production pages for leftover staging restrictions. Those checks show that the site works; they do not show that search engines will index or rank it.