Compatibility
Technical dependency
A feature requires specific versions of WordPress, PHP, a theme, a builder or other extensions. Changing one layer can create errors elsewhere.
Guides
A CMS can do almost anything. The decisive question arrives when something breaks: who can fix it immediately, and how many vendors, versions, licences and release schedules does your website depend on?
Direct answer
WordPress is not a problem because it uses plugins: modularity is one of its strengths. Risk appears when a business function depends on a chain that nobody governs end to end. An update may be necessary for security yet incompatible with a theme, builder or add-on; editing vendor files exposes the patch to overwriting; waiting for a fix hands priorities and timing to someone else. A bespoke website does not eliminate every dependency: it reduces them, makes them explicit and concentrates responsibility on the project’s code.
Often yes at the level of the visible feature. Not necessarily at the level of control, responsibility and ability to intervene.
The interface may look unified, but the output depends on different software, organisations and release calendars.
It manages content, users, routing and base APIs. Its evolution and requirements can change how extensions behave.
It controls presentation and often functionality. If abandoned or heavily customised, every update becomes a delicate decision.
It adds an editor, components, CSS, JavaScript and its own data models. It may require a specific core or Pro version.
SEO, forms, caching, security, languages, ecommerce and cookies introduce code and independent update cycles.
An extension can depend on another extension, not just WordPress. Compatibility becomes a matrix rather than a line.
Versions, memory, modules, caching and server configuration can make previously working code incompatible.
Updates and premium services may depend on renewals, accounts, external infrastructure and support priorities.
The technical problem becomes commercial when it interrupts leads, sales, bookings, indexing or staff activity.
The dependency that matters
Risk is not measured by counting plugins. It is measured by asking which features are critical, who controls their code, how long recovery takes and what happens if a vendor stops cooperating.
Two websites with the same number of plugins can carry completely different risks.
Compatibility
A feature requires specific versions of WordPress, PHP, a theme, a builder or other extensions. Changing one layer can create errors elsewhere.
Risk window
A vulnerability requires a vendor fix. Until it arrives, you must mitigate, disable, replace or accept the risk.
Intervention time
The person supporting the website may diagnose the problem without controlling the roadmap of the component that caused it.
Contracts and continuity
Expired licences, changed plans, usage limits, tiered support or discontinued products alter the cost and availability of a feature.
Composite vendor lock-in
WordPress and many plugin codebases are available. In theory they can be studied, changed or forked. In practice, maintaining a secure fork requires skills, tests and continuous updates.
Licence freedom reduces some legal and technical constraints; it does not remove the cost of taking over the work previously done by the vendor.
Delaying updates increases exposure; applying them without inventory, staging and recovery can interrupt the website.
WordPress recommends keeping plugins and themes updated. That is sensible: new versions fix vulnerabilities, incompatibilities and defects.
The same ecosystem also requires backups and rollback capability because an update can fail or conflict with other components. Automation does not remove that responsibility.
Correct maintenance is therefore a process: know the dependencies, read relevant changes, test in a controlled environment, verify critical paths and be able to return to a working version.
“Update everything” is not an operating plan. It is an action that requires testing, observation and recovery.
Versions, licences, declared compatibility, data, files and the recovery procedure must be known.
Login, editor, forms, checkout, caching, SEO, languages and integrations must be tested beyond the homepage.
PHP and JavaScript errors, HTTP status, performance and conversions must confirm the website is more than merely visible.
WordPress documentation recommends backups and recovery because automatic updates can go wrong: read the official procedure.
An urgent patch can bring the website back. The question is who will maintain that deviation at the next release.
| Intervention | Initial speed | Update effect | Later debt |
|---|---|---|---|
| Edit vendor files | May be immediate | May be overwritten | The patch must be remembered and reapplied or abandoned |
| Official hook, filter or API | Requires a suitable extension point | More resilient, not guaranteed forever | Must be tested against supported versions |
| Child theme | Suitable for expected customisation | Protects separated files | Does not repair defects inside a theme, builder or plugin |
| Independently maintained fork | Slow to establish correctly | Full control, manual updates | Security, compatibility and merging become your responsibility |
This is not a client story or an invented scenario: it is an error and recovery path described by Elementor’s official support.
Version dependency
Elementor documents a specific fatal error that can appear when Elementor Core is updated but Elementor Pro is not. The website may go down and the administration area may be unavailable.
The stated solution is to remove or disable Elementor Pro through FTP, regain access and install an updated compatible version. For other post-update issues, Elementor also documents rollback.
One component changes version while the other remains behind.
The Pro component calls a method that the installed combination does not expose as expected.
This is not a visual detail: it can prevent both the website and WP Admin from loading.
The official procedure requires FTP or a file manager to disable the component blocking administration.
A compatible release is installed, or the system returns to an earlier version after a backup has been prepared.
Restoring WP Admin does not prove that forms, templates, checkout, caching and the frontend are all correct.
Case reconstructed from the official procedure, updated on 18 May 2026: fatal error after installation or update.
On more than one occasion I found JavaScript errors, prepared a diagnosis and urgently contacted the vendor. The operational answer was to wait for a later release.
The point is not to blame Yoast: all software contains bugs and every vendor must assess impact, priority, compatibility and release timing. Support can acknowledge a problem without being able to deliver an immediate fix.
From the company’s perspective, however, the outcome is unchanged: the person managing the website knows the cause but does not control the fix calendar. A patch to plugin files may solve the urgency while creating a change destined to be overwritten.
That is the difference between diagnostic expertise and operational control. You can know exactly what is wrong and still depend on another team’s decision.
Error reproduced, console and conditions identified: the problem is no longer generic.
The vendor receives useful information and can confirm that the defect belongs to the component.
The fix enters the vendor’s roadmap, not that of the business suffering the problem.
A serious plan separates containment, recovery and the permanent correction.
Time, latest change, logs, stack trace, console, HTTP status, affected pages and users.
If compromise is possible, preserve logs, restrict access and do not confuse recovery with destroying evidence.
Disable, isolate or restrict only what causes harm, while assessing the effect on business functions.
File and data rollback must return the system to a verified state, not simply to a screen that opens.
Vendor, integrator or internal team need explicit responsibilities and timings; “we opened a ticket” is not recovery.
Leads, sales, login, editor, payments, email, indexing and caching must be verified according to the website’s role.
Real recovery time
The value of technology is not visible only when it works. It appears in how quickly an accountable person can understand the failure, intervene without side effects and prove that service integrity has been restored.
PHP, servers, browsers, web standards, selected libraries and external services still exist. The difference lies in scope and responsibility.
Control, not isolation
In my method every layer must justify its presence. Common features enter the GLP CORE as versioned components; specific features are integrated into the project without turning every requirement into an independent plugin.
This table does not declare a universal winner. It shows where responsibility sits when the website must change or recover from a problem.
| Situation | CMS ecosystem | Controlled bespoke | Business question |
|---|---|---|---|
| New feature | Ready-made plugin or extension combination | Feature designed within the existing scope | Do you need initial speed or long-term coherence? |
| Urgent error | Diagnosis distributed across several vendors | Responsibility concentrated on the project | Who can release the fix today? |
| Update | Matrix across core, theme, plugins and runtime | Declared dependencies and tests of affected code | Do tested staging and rollback exist? |
| External vulnerability | Wait for patch, mitigate or replace | Direct fix when owned code is affected | How long is the exposure window? |
| Licence or service discontinued | Feature, updates or support may change | Application code remains in the project | What stops working without renewal? |
| Technology change | Content may export; the system often needs rebuilding | Portability depends on documentation and standards used | Do you own data, files, accounts and the URL map? |
To choose the technology, also read WordPress or a bespoke website ; for attack surface, see the website security guide.
The total includes what you pay, what you must maintain and what you lose when the website does not do its job.
Hosting, theme, licences, renewals, premium support, backups, security, staging and supporting tools.
Updates, tests, incompatibilities, logs, tickets, renewals and coordination among supplier, host and vendors.
Urgent diagnosis, downtime, recovery, data restoration, lost orders and out-of-hours work.
Delayed features, slowed campaigns, insufficient performance and business decisions adapted to platform constraints.
TCO, not entry price
An inexpensive plugin can be an excellent purchase. It becomes costly when the feature is critical and the business has neither control over the code nor a practical exit path.
To estimate licences, maintenance and time over five years, use the WordPress total-cost calculator.
Recognising suitable cases makes the opposite choice more credible.
A large editorial team needs roles, reviews, scheduling and day-to-day publishing autonomy.
An established extension can provide more value than a rewrite when requirements and compromises genuinely match.
To validate a temporary idea, speed and initial cost may matter more than long-term system ownership.
Inventory, staging, updates, security, backups and incident response turn the CMS from a shortcut into managed infrastructure.
Bespoke development gains value when the website is an operational asset intended to last and differentiate the business.
Leads, bookings, catalogue or checkout make downtime and regressions a concrete business risk.
Structure, rendering and weight should follow real content rather than compensate for a general-purpose stack.
APIs, flows, data and permissions should enter a coherent architecture instead of crossing several bridge plugins.
Reducing logins, exposed databases, scripts, endpoints and third parties may be more effective than adding corrective layers.
Features and content should change without chasing a theme, builder or commercial plan.
The business wants to know who owns the full output and can intervene, not which vendor will open the next ticket.
The decision in one sentence
I design a deliberate reduction of the dependency surface and take direct responsibility for the code governing the website.
The difference is not being able to add a feature. It is being able to understand, correct and verify it without waiting for a chain of vendor roadmaps.
If many answers are unknown, the problem is not necessarily WordPress: it is the absence of technical governance.
Core, theme, child theme, builder, plugins, add-ons, snippets, APIs, licences and server versions.
Components do not have equal impact on sales, leads, data, SEO and internal activity.
Domain, hosting, licences, email accounts and services should belong to the business or be transferable.
The test copy must represent enough configuration, data and journeys to expose real regressions.
An automatically generated file is not a strategy until you know it can restore website and data.
Every change should live in extension points, child themes or controlled documented components.
Contacts, responsibilities, access, priorities and authority to disable a feature should be established in advance.
A non-critical feature may wait; checkout, forms and vulnerabilities may need an immediate alternative path.
Content, media, orders, users, metadata and the address map are required to change systems without losing the digital asset.
Add licences, hosting, maintenance, incidents, internal hours, rebuilds and lost opportunities—not only the initial quotation.
I do not patch the old stack forever: I record what matters and rebuild the outcome without carrying over the technical debt.
GLP method
A WordPress migration starts with an audit and separates useful content, URLs, data and features from the theme, plugins and configurations that created fragility.
| Path | Scope | Price before tax |
|---|---|---|
| Migration audit | Analysis of the current WordPress website, dependencies, content, URLs, risks and action plan | €175 |
| Essential | Compact rebuild with up to four independent pages | €1,000 |
| Pro | Architecture, pages and features for a more substantial business website | from €2,500 |
| Essential e-commerce | Focused catalogue, product pages, basket and checkout without WooCommerce | from €4,500 |
For process, conditions and frequently asked questions, see the dedicated WordPress migration page. The final quotation depends on the content, integrations and actual approved scope.
Vendor and WordPress project documentation. The Elementor case is an official procedure; the Yoast episode is explicitly identified as personal professional experience.
Declared or untested compatibility, updates and the varied nature of available extensions.
Update configuration, backups and recovery capability when a release causes problems.
Disabling plugins through FTP and overwriting direct changes during updates.
Core class that reads requirements, dependants, unmet dependencies and circular dependencies.
The intended method for separating theme customisation and reducing the risk of losing it during updates.
Website unavailable, Core and Pro incompatibility, recovery through FTP and version realignment.
Restoring an earlier release when an update causes problems and the need for a prior backup.
A recent example of console, integration and security issues corrected through a later release.
Backups, core files involved in an upgrade and the manual procedure when automatic updating fails.
Sources checked on 7 August 2026. Versions, procedures and commercial terms change: always verify the current documentation for the component actually installed before acting.
Direct answers about updates, patches, security, ownership, migration and bespoke websites.
No. It can suit frequent editorial work, standard projects and organisations able to govern updates, compatibility, backups and security. It becomes risky when treated as a system that maintains itself.
No. Quality, exposure, configuration, maintenance and the value of affected features matter. Reducing components helps, but does not replace correct development and management.
Updates reduce known risks but can introduce incompatibilities or regressions. Restorable backups, proportionate staging, tests of critical journeys and post-release monitoring are required.
Technically, often yes. If you edit vendor-distributed files, the next update may overwrite the patch. Official hooks and APIs are preferable; a fork means taking on future maintenance and security.
No. It protects correctly placed theme customisation, but does not automatically fix defects inside plugins, builders, WordPress or external services.
Code availability provides important freedom, but operational independence requires the skills and budget to maintain changes. Data, editors, shortcodes, builders and licences can still create exit costs.
Elementor documents cases in which a fatal error prevents access to WP Admin. Recovery involves disabling or removing the component through FTP and realigning versions.
No. It depends at least on its language, server, browsers, standards and selected services. The professional objective is to reduce avoidable dependencies, document necessary ones and directly control application code.
If you do not know the components, licences, account owners, compatibility, recovery procedure, incident owner and ability to export data and URLs, technical governance is missing regardless of plugin count.
No, when the migration inventories content and URLs, preserves what creates value, sets correct redirects and verifies canonicals, sitemaps, metadata, structured data and performance before and after launch.
The migration audit costs €175 before tax and analyses the website, dependencies, content, URLs, risks and rebuild path. The new website quotation is defined against the approved scope.
Current packages start with Essential at €1,000, Pro from €2,500 and Essential e-commerce from €4,500, before tax. Size, content, features, languages and integrations determine the final cost.
Next step
The migration audit identifies features, content, URLs, risks and costs in the current chain. You can then decide with concrete evidence whether to keep it or rebuild the website to measure.