Before changing anything, work out which of these you actually have, because they have different causes and different fixes:
- The plugin will not activate
- It activates but nothing happens
- It worked yesterday and stopped today
- It works for you but not for visitors
- Activating it takes the whole site down
Most articles on this give you one list of fixes to try in order. That wastes time, because the fix for a caching problem will never solve a PHP version problem. Narrow it first.
The two minute checks
Do these before anything that involves a file.
Clear every cache, then hard refresh. On a WordPress site there are usually three or four stacked: a caching plugin, the host’s server cache, a CDN, and your browser. Clearing one and seeing no change proves nothing. Clear all of them, then reload with the cache bypassed.
Check in a private window. If the plugin behaves correctly there and not in your normal window, the problem is your browser or your login session, not the plugin.
Check as a logged out visitor. The reverse case matters more. A plugin that works while you are logged in and fails for everyone else is almost always caching, because most caching setups serve a stored page to logged out visitors and bypass the cache for administrators. You have been looking at the live site. Everyone else has been looking at a photograph of it.
Look at the plugin’s own settings screen. A surprising number of plugins install deactivated in all but name, waiting for a licence key, an API connection, or a single enable toggle.
Read the error instead of guessing
This is the step people skip, and it is the one that usually ends the problem in a minute.
WordPress can log every PHP error to a file. Add this to wp-config.php, and it has to go before the line that reads that is all, stop editing:
define( 'WP_DEBUG', true );
define( 'WP_DEBUG_LOG', true );
define( 'WP_DEBUG_DISPLAY', false );
@ini_set( 'display_errors', 0 );That writes everything to wp-content/debug.log while hiding the errors from the page, so visitors see nothing unusual. Reproduce the problem, then open the log and read the last entries. The file name and line number in a fatal error tell you which plugin is responsible without any deactivating at all.
Three things worth knowing about this:
- The WordPress documentation is explicit that debugging tools are meant for local and staging installs rather than live sites. Turn it on, get your answer, turn it off
- Do not put the values in quotes.
define( 'WP_DEBUG', 'false' )is a non-empty string, which PHP reads as true, so you switch debugging on while believing you switched it off - Back the site up, or work on staging, before editing wp-config.php at all
If you cannot edit files, the host’s control panel usually has an error log of its own, and the Tools, Site Health screen inside WordPress reports a good deal without any file access.
Isolating a conflict without taking the site down
The standard advice is to deactivate every plugin, confirm the problem goes away, then reactivate one at a time. It works, and on a live site it is a terrible idea, because for the duration your visitors get a site with no forms, no shop and no security plugin.
There are two better versions.
Troubleshooting mode. The Health Check and Troubleshooting plugin from the WordPress team disables plugins and switches the theme for your session only. Everyone else keeps seeing the normal site. You can then re-enable plugins one by one from the admin bar and watch for the moment the problem returns. This is the right tool for a site that is live.
Staging. If your host offers one click staging, clone the site and break the copy instead. Most managed hosts and many cPanel hosts include this now.
Whichever route you take, test the theme too. A conflict between a plugin and a theme is as common as one between two plugins, and switching temporarily to a default theme rules half the possibilities out in one step.
When you cannot reach wp-admin at all
If activating the plugin produced a white screen or a critical error message and you are locked out, you have two ways back in.
Recovery mode. Since WordPress 5.2 a fatal error triggers an email to the site administrator address containing a secret link into recovery mode, where the offending plugin is paused and you can get to the dashboard. Check that inbox, including spam, before doing anything drastic. Two caveats: the email is rate limited, by default to one per hour, and the handler does not catch errors that happen while must use plugins are loading, so an mu-plugin fault produces no email at all.
Rename the folder. Over FTP, SFTP or the host’s file manager, go to wp-content/plugins/ and rename the suspect plugin’s folder, adding anything to the end. WordPress cannot find it, so it deactivates it and lets you back in. Renaming the whole plugins folder does the same for all of them at once.
Rename it back afterwards and the plugin returns, deactivated, with its settings intact. This is a reversible operation, which is why it is the standard move rather than deleting.
Now match the symptom to the cause
It will not activate
Usually an environment problem rather than a bug.
- PHP version. Modern plugins declare a minimum, and an older PHP will refuse or fatal. Site Health reports your version
- WordPress version. The same applies in both directions, and a plugin written for a much older WordPress can break on a current one
- Missing PHP extension. Image, zip and encryption features need extensions your host may not have enabled
- Memory. Activation runs installation routines, which is the moment a tight memory limit shows up
It activates but does nothing
Start with the obvious and work outward. Is there a setting that enables it? Is the shortcode or block actually on the page? Is the feature meant to appear somewhere you have not looked?
After that, suspect JavaScript. Open the browser console and reload. Script errors from one plugin will stop later scripts from another running at all, which produces exactly this symptom: a plugin that is installed, active, correctly configured and completely inert. Minification and combining scripts, usually from a performance plugin, is a frequent cause, so switch those options off and retest.
It worked yesterday
Something changed, and there are only so many candidates. In rough order of likelihood: the plugin updated, WordPress updated, another plugin updated, the theme updated, your host changed the PHP version, or a licence expired.
Automatic updates make this harder to see, because nobody remembers doing anything. The site activity or update log will usually tell you what moved and when, and matching that timestamp against when the problem started is faster than any amount of deactivating.
If the plugin itself updated and broke, most plugins can be rolled back to a previous version from the WordPress.org repository, which buys you time to report the bug rather than forcing a same-day replacement.
It works for you but not for visitors
Three causes, in order of how often they turn out to be right.
Caching, as described above. Exclude the relevant pages from the cache and retest.
Roles and capabilities. You are an administrator and can do things other users cannot. Create a test account at the role your visitors have and look through their eyes.
A firewall or security rule. This one is underdiagnosed. Security plugins and server level rules such as mod_security block requests that look suspicious, and plugin requests to admin-ajax or the REST API often qualify. The giveaway is a request that returns 403 in the browser network tab while everything else loads normally. Your host can check whether a rule fired and whitelist it.
Activating it takes the site down
A fatal error, which means a PHP conflict rather than a configuration mistake. Use recovery mode or the folder rename above to get back in, then read debug.log for the file and line.
The commonest cause is two plugins loading different versions of the same library, which produces an error about a class or function already being declared. Neither plugin is broken. They simply cannot coexist, and the answer is replacing one of them rather than fixing either.
The settings will not save
A distinct symptom with its own causes, and the interesting one is almost never mentioned.
max_input_vars. PHP limits how many variables a single form submission may contain, commonly to 1000. A large settings page, a long menu or a theme options panel can exceed that, and PHP then silently discards everything past the limit. Nothing errors. The page reloads looking successful and half your settings are gone, usually the ones furthest down the form. Raising the limit is a PHP setting, so it is a host question.
An expired nonce. If the form sat open in a tab overnight, its security token has expired and the submission is rejected. Reload the page and fill it in again.
A firewall blocking the POST. Server rules are more suspicious of submissions than of page loads, especially where a field contains code, a URL or a snippet of HTML. A settings page that saves everything except the one field containing a script tag is this, every time.
Database writes failing. Rarer, but a full disk or a corrupted options table produces it. The host can confirm disk space in seconds.
The white screen
A blank page with no message is PHP dying without being allowed to say why, which is why the first move is always making it talk. Enable debug logging and reload, and the blank page becomes a specific error in debug.log.
Three things produce most of them. A fatal PHP error, usually a conflict or a version mismatch. Exhausted memory, which the log reports as an allowed memory size message naming the file that asked for the last allocation, though that file is often innocent and simply unlucky. And an exhausted execution time on a long running task.
If the front of the site is white but /wp-admin still loads, the fault is in a theme template or a front end only plugin. If both are white, it loads earlier than that, which points at a plugin, an mu-plugin or wp-config.php itself.
The install or upload fails
- Upload size. A plugin zip larger than the upload limit is rejected. Site Health shows the limit, and installing from the repository rather than by upload sidesteps it
- Permissions. A could not create directory message means WordPress cannot write to wp-content/plugins. That is ownership rather than a WordPress problem
- Disk full. Produces failures that look like anything else. Worth ruling out first because it takes one glance
- Update failed, another update is in progress. A previous update died partway and left a lock behind. It clears on its own after a while
The host side causes
Worth knowing because no amount of work inside WordPress will touch them.
PHP memory limit. Exhausting it produces either a white screen or an allowed memory size error in the log. Raising it is usually a host setting rather than something you can do from WordPress.
Execution time. Import, backup and optimisation plugins run long tasks that a short max execution time kills partway through, which looks like the plugin silently failing.
Server firewall rules. Covered above. Ask the host to check their logs for your IP at the time the request failed, which is far more productive than asking whether anything is blocked.
Object cache. Where a persistent object cache is in use, stale cached options can keep producing the old behaviour after you have changed a setting. Flushing it is part of clearing caches, and it is the one people forget.
File permissions. A plugin that cannot write where it expects to will fail at install or when saving, often with no visible error.
Reading the network tab
Worth a section of its own, because for anything interactive, a form, a filter, a cart, an admin screen that spins forever, this is where the answer is and most people never open it.
Open the browser developer tools, go to the network tab, reload, then do the thing that fails. Watch the requests that appear, and look at the status code on each.
| What you see | What it usually means |
|---|---|
| 403 on admin-ajax.php or a wp-json route | A security plugin or server firewall rule is blocking it. Ask the host to check which rule fired |
| 404 on a wp-json route | The REST API is unreachable or the route is not registered. Permalinks, a disabled REST API, or the plugin failing to load |
| 500 | A PHP fatal error inside the request. The detail is in debug.log |
| 200 with an empty or truncated response | PHP died quietly, often memory or execution time. Check the log |
| 200 with HTML where JSON was expected | Something printed output before the response, commonly a PHP notice or a stray blank line in a plugin or theme file |
| The request never appears at all | JavaScript failed before sending it. The console tab has the error |
That last row is the one that saves the most time. If no request is made, nothing on the server is at fault and there is no point looking there. The problem is a script error, and the console will name the file.
The 200 with HTML case is worth recognising too, because it looks successful. A plugin expecting JSON receives a page of markup, cannot parse it, and gives up silently. The cause is usually something printing a warning that WP_DEBUG_DISPLAY would have revealed.
Before you contact support
A ticket that contains the following gets answered properly first time instead of producing a reply asking for it.
- Exactly what you expected to happen and what happened instead
- The relevant lines from debug.log, not a description of them
- Your WordPress version, PHP version and the plugin version
- What you have already ruled out, specifically that you tested with other plugins disabled and with a default theme
- Whether it affects logged out visitors, logged in users, or both
- When it started and what changed around then
Developers close a lot of tickets as cannot reproduce. Most of those are conflicts the reporter never isolated, so the half hour you spend on the steps above is the difference between a fix and a fortnight.
Stopping it happening again
None of this is exotic and all of it is dull, which is why it does not get done.
Keep a staging copy and apply updates there first. Take a backup before updating, and make sure you have actually restored one at some point, because an untested backup is a hope rather than a plan. Update in small batches so that when something breaks you know what did it. Remove plugins you are not using rather than leaving them deactivated, since an abandoned plugin is a security problem whether it is running or not.
And keep an eye on what is no longer maintained. A plugin that has not been updated in years will eventually meet a PHP version it cannot survive, and the failure will arrive on the host’s schedule rather than yours. We set out the wider routine in the WordPress maintenance checklist.
The order to work through
- Clear every cache and retest, logged in and logged out
- Turn on debug logging and read the log
- Use troubleshooting mode to test with other plugins off and a default theme, without affecting visitors
- If locked out, use the recovery mode email or rename the plugin folder
- Check PHP version, WordPress version and memory in Site Health
- Check the browser console and network tab for script errors and 403s
- Then, and only then, open a support ticket with the evidence
Steps one and two solve most of these. The rest exist for the ones that do not.
Common questions
Why is my WordPress plugin not working after an update?
Something in the environment changed. The usual candidates are the plugin updating, WordPress updating, another plugin updating, the theme updating, the host changing the PHP version, or a licence expiring. Check the update log for what moved and match the timestamp against when the problem started.
How do I find which plugin is causing the problem?
Read wp-content/debug.log after enabling WP_DEBUG and WP_DEBUG_LOG, because a fatal error names the file and line. If that does not identify it, use the Health Check and Troubleshooting plugin, which disables plugins for your session only so visitors are unaffected, and re-enable them one at a time.
What do I do if a plugin locks me out of wp-admin?
Check the site administrator email for a WordPress recovery mode link, which is sent automatically on a fatal error and pauses the offending plugin. Otherwise rename that plugin folder inside wp-content/plugins over FTP or the host file manager, which deactivates it. Renaming it back restores the plugin with its settings.
Why does the plugin work for me but not for visitors?
Most often caching, because caching setups usually serve a stored page to logged out visitors while bypassing the cache for administrators. The other two causes are user role capabilities, and a security plugin or server firewall rule blocking the plugin requests, which shows as a 403 in the browser network tab.
Why do my plugin settings not save?
The least known cause is PHP max_input_vars, commonly limited to 1000. A large settings form exceeding it has the extra fields silently discarded with no error, so the page looks like it saved. The other causes are an expired security nonce from a form left open too long, a server firewall rejecting the submission, and database writes failing because the disk is full.
Is it safe to turn on WP_DEBUG on a live site?
The WordPress documentation says the debug tools are meant for local and staging installs rather than live sites. If you must use it on a live site, set WP_DEBUG_DISPLAY to false and display_errors to 0 so errors go to the log rather than the page, take a backup first, and turn it off again as soon as you have your answer.
A plugin activates but nothing happens. What now?
Check whether it needs enabling, a licence key or an API connection, and whether the block or shortcode is actually on the page. Then open the browser console, because a script error from one plugin stops later scripts running. Minifying or combining JavaScript is a frequent cause, so disable those options and retest.