When Not to Migrate: Sites That Should Stay Where They Are
Some sites should not move. Sites that depend on an app, a membership, a large store or a plugin with no equivalent lose more than they gain. The questions that tell you to stay, and the smaller moves that get most of the benefit without the risk.
A site should stay where it is when the thing it does best is something the new platform cannot do, when the move would cost more search traffic than it saves in maintenance, or when the reason for moving is a feeling about the old platform rather than a problem with it. An AI agent makes the mechanical part of a migration cheap. It does not make the decision cheap, and a cheap build is the wrong reason to make an expensive move.
This post is part of our guide to Migrating a Website With AI.
The site is an app
A website builder builds websites: pages, posts, products, forms. If your site has user accounts, a members-only area, a custom database, a booking engine with its own logic, or a dashboard people log into, it is an app with a website around it. Beste, for one, does not do user accounts, custom databases or app logic, and most hosted builders draw a similar line.
The test: list every feature on the old site and mark the ones that need a login or store data per visitor. If those features are why people come, the site stays. If they are a bolt-on that nobody used, the site can move and the feature can be dropped. Be honest about which. A membership plugin with four members should not keep a 300-page site where it is; a booking system that is the whole business should.
A middle path exists: move the marketing site and the blog, keep the app on its own subdomain. The redirect map covers the pages that moved, the app keeps its URLs, and both get the attention they need.
The store is large or complex
A catalog with options and variants, a cart that links to a checkout, and product structured data cover most small stores. They do not cover a store with thousands of SKUs, inventory synced from a warehouse, multi-currency tax rules, subscriptions, or a fulfillment integration. Wix's and Squarespace's own export tools cap product exports at 5,000 and 10,000 rows per file, per their Wix Stores and Squarespace help pages, which gives a sense of where "large" begins.
If the store is the business and it is large, the store stays on a commerce platform. The content around it can still move, and the catalog on the new site can link its checkout to the commerce platform.
The old URLs are worth more than the new design
Google's site move guidance is clear that a move with URL changes carries risk even when done well: temporary ranking fluctuations, weeks or more for a medium-sized site to settle, and redirects to maintain for at least a year. A site with hundreds of ranking pages, deep external links and a URL structure the new platform cannot reproduce is paying that cost in full.
Three questions decide it:
- Can the new platform reproduce most of the old paths? If yes, most rows in the redirect map disappear and the risk falls with them. If the old site uses date-based permalinks or deep category paths that cannot be matched, every page moves and every page is at some risk.
- How much of the traffic comes from search to old URLs? Export it from Search Console. If it is most of the business, plan the move as Redirect Mapping From a CSV describes and accept a careful month. If it is a small share, the risk is small too.
- Is the site mostly evergreen reference pages? Those are the pages that lose the most when a redirect misses, because they rank on age and links. A site of news posts that stop earning traffic after a month has little to lose.
A plugin does the one thing the business depends on
A WordPress site with an event calendar that sells tickets, a directory plugin with 2,000 listings, a learning plugin with courses and progress, or a real-estate plugin pulling from a listing feed is not a WordPress site with content. It is a plugin with a site around it. Migrating From WordPress After PHP2Shell has a table of common plugins and what replaces each; the ones with no replacement are the ones that keep a site where it is. If the plugin's job has no equivalent on the new platform and is central, stay.
The opposite case is common too: ten plugins installed, two in use, both replaceable by a section. That is a site that should move, and the plugin count was the reason it felt heavy.
The problem is maintenance, not the platform
Many requests to migrate are really maintenance complaints: updates that broke things, a host that went slow, a designer who left. A migration fixes those by replacing everything, which works but is the most expensive fix. Before moving, ask whether a smaller change solves it:
- Content is stale. That is a routine, not a rebuild. Keeping a Website Current With an AI Agent is about this.
- The site is slow. Often a hosting or image problem on the same platform.
- Nobody can edit it. Sometimes a training problem, sometimes a sign that the editor is the issue and a move is right.
- The design is dated. A new theme on the same platform keeps every URL.
- A security alert about the server software. The PHP2Shell release on 21 October 2026 is the current example. The fix was a PHP update from the host, and the WordPress.org announcement says no WordPress update is needed. That is not a reason to move. It becomes one only when the site cannot run on a supported PHP version because something on it breaks and nobody can fix it; the PHP2Shell section of the WordPress post covers how to tell.
If the answer to each is a smaller change, make the smaller change. If two or three of them point the same way, the move is probably right, and you have a list of what the new site must do better.
The site is about to be replaced anyway
A business that is rebranding, merging, changing its name or moving its domain in six months should not migrate now and again then. Google's site move guide advises planning changes one after the other rather than all at once; two moves in a year is the opposite. Wait and do both changes as one planned move, with one redirect map.
A one-page test
Answer each line for your own site. It is not a formula, but if most answers land in the right-hand column, the move is worth planning.
| Question | Points to staying | Points to moving |
|---|---|---|
| What do visitors come for? | Something they log into, book or manage | Information, products, contact |
| How many plugins or apps do real work? | One or more with no equivalent elsewhere | None, or only ones a builder replaces |
| Can the new platform keep most old paths? | No, almost every URL changes | Yes, most stay the same |
| Who keeps the server and software current? | A host or developer who does it reliably | Nobody, or it keeps breaking |
| Can the site run on supported PHP today? | Yes, with a routine upgrade | No, something pins it to an old version |
| Is a rebrand or domain change coming soon? | Yes, within six months | No |
| Who edits the content? | A developer, and that works | People who avoid the editor |
The two rows in the middle carry the most weight this autumn. A site that nobody keeps current and that cannot run on a supported PHP version is a site whose risk grows every month, and the PHP2Shell release made that visible for many owners for the first time.
If you do stay
Staying is a decision, not a default. Write down why, and what would change the answer. Then do the parts of a migration that are useful anyway: build the content inventory, because it shows what the site has; clean up the pages that are dead; fix the internal links; and make sure someone is named as responsible for updating the server software. Content Inventory With an AI Agent describes the inventory, and it is the same document whether the site moves or not. If the answer changes next year, the hard part is already done.
How this works in Beste
Beste builds websites, not apps: pages, posts, products, forms and the files around them. So it is worth saying plainly what it does not do, so a decision does not rest on a guess:
- No user accounts, members areas, custom databases or app logic.
- No payment processing. The products catalog handles products with options and variants and sells through a Stripe payment link, a quote, WhatsApp or a link to Shopify, which is the right size for a small catalog and not a replacement for a large commerce setup.
- No importer for Wix or Squarespace; WordPress posts can be imported from the Posts panel in Studio with their images and slugs, and the rest is recreated.
- No exportable codebase and no arbitrary markup, per the no code to maintain page. Pages and posts can be read as Markdown and images downloaded, so the content can always leave.
- No form response export and no post scheduling yet.
What it does take off your hands is the maintenance side: hosting, HTTPS and platform updates are Beste's job, so there is no server, plugin or PHP version on your side to keep current. For a site that does move, nested pages reproduce most old paths exactly, and the ones that change go into the CSV redirect import on Pro described on the technical SEO page; keep Pro while those redirects matter, because they stop working if the plan is cancelled.
If you are not sure, test before deciding. The Free plan has no time limit and you can build on a beste.co subdomain first, so an agent can rebuild a slice of the site, say the ten pages that earn the most traffic, as a draft you can look at next to the old one. That changes nothing on the live site and answers the question better than a feature list.
Sources
Checked on 23 October 2026.
