Perfmatters vs a Caching Plugin: Only One Feature Does Not Overlap

Updated October 2026

Perfmatters costs $29.95 a year for one site, confirmed against its own pricing page on 25 October 2026, and it does not cache a single page. LiteSpeed Cache is free, reports 7,000,000+ active installs through the wordpress.org Plugin API on the same day, and has no way to turn one plugin’s JavaScript off on a single URL. The comparison people expect is which of the two makes a WordPress site faster. The more useful comparison is which moment of the request each one acts on, because that is what decides whether you need one, the other, or both. This article maps those moments, measures how much of the feature list genuinely overlaps, and gives you a census you can run on your own site before you spend anything. Updated on 25 October 2026: FlyingPress added as a third cache, every price and quote re-read that day.

Quick Summary: What to Do Before You Spend Anything

Work down this table until a row describes your site. The criteria come first on purpose. A product name only means something once you know which question you are asking.

Two things are true at once and most articles about this pick only one of them. Perfmatters duplicates a large part of what a modern caching plugin already does. And it holds one capability that no caching plugin in this field ships at all.

Four Moments in a WordPress Request, and What Acts at Each

A WordPress page request has four points where software can intervene. Every performance plugin you have ever installed acts on at least one of them, and almost none acts on all four. Sorting the tools by moment is what turns a feature grid into a decision.

Moment one, before PHP starts. A page cache has already written this URL’s HTML to a file. The web server hands that file back and PHP never boots. This is where the response time saving lives, and it is the only moment where a saving is measured in server milliseconds.

Moment two, PHP boots and plugins load. WordPress reads the active plugin list and includes every one of them. A plugin that runs here costs PHP execution and database queries whether or not this URL uses it. Very few tools can intervene at this moment, because doing so means changing what WordPress loads before WordPress has finished loading.

Moment three, PHP builds the page. Themes and plugins enqueue stylesheets and scripts. A contact form plugin adds its JavaScript to every page, not only the contact page. This is the moment a script manager acts on, and it is the moment that decides how many files the browser will be asked to fetch.

Moment four, the HTML is finished. Now the markup can be rewritten: minified, scripts deferred or delayed, images lazy-loaded, unused CSS stripped. Both categories of tool act here, which is exactly where the duplication comes from.

There is a detail here that changes the whole argument, and it is the reason these two tools compose rather than compete. Moments two and three only execute on a cache miss. But their result is baked into the HTML file the cache then stores, so every cache hit afterwards serves the leaner page. Unloading a script does not stop working because the page is cached. It gets frozen into the cached copy.

The reverse is not true. Moment one’s saving stops at the server response. When we pulled our own scan archive on 10 September 2026, across 413 WordPress sites the median server response accounted for 5.9% of the wait before the largest element painted. A perfect cache still leaves the browser downloading and parsing everything moments three and four put on the page.

The Overlap Is Most of the Feature List

Read Perfmatters’ feature page next to any modern caching plugin’s and the duplication is not subtle. Minify JavaScript and CSS. Defer JavaScript. Delay JavaScript until interaction. Remove unused CSS. Lazy load images, iframes and video, with a facade for YouTube. Preload, preconnect, DNS prefetch, fetch priority. CDN URL rewriting. Database cleanup with scheduling. Heartbeat control. Disabling emoji, embeds, dashicons and XML-RPC. Every one of those appears in both product categories.

You do not have to take our word for the overlap, because Perfmatters documents it themselves and tells you how to resolve it. Their guide to running alongside LiteSpeed Cache states: “we usually recommend turning everything off in LiteSpeed Cache except for the cache features and letting Perfmatters handle all the front-end code optimizations.” The same page adds the general rule: “The important thing is simply not to enable the same feature in both plugins.” Their WP Fastest Cache guide does the same job for a different plugin, listing which cache settings to leave on and which to hand over.

That is a vendor being honest about its own scope, and it is the most useful sentence either company has published on this question. It also frames the purchase accurately. Buying Perfmatters and running it as intended means switching off most of your caching plugin’s optimization panel and running a second implementation of it instead.

A tick is included, a coin is paid tier only, a cross is not available, an em-rule means undocumented. Cache marks come from our own 80-capability comparison, which records the tier each feature ships in. Perfmatters and FlyingPress marks come from their own pages, re-read 25 October 2026. Every FlyingPress row is paid because it has no free tier, and its analytics cross is because the eight domains it self-hosts by default are CDN libraries.

The Script Manager: The One Thing No Caching Plugin Ships

Strip out everything that overlaps and one row is left standing. Perfmatters’ Script Manager lists every stylesheet and script loading on the page you are looking at, grouped by the plugin or theme that enqueued it, and lets you switch each one off. Not globally, which any exclusion field can do, but per URL, per post type, per device, and per logged-in state, with exceptions layered on top and regular expressions available for the awkward cases.

The canonical example is in Perfmatters’ own documentation: a contact form plugin loads its CSS and JavaScript on every page of the site, including the ones with no form on them. Disable it everywhere, add an exception for the contact page, and the rest of the site stops fetching it. Their docs claim that on their own sites this “instantly decreased our HTTP requests on our homepage by over 10.”

MU Mode Reaches Past Assets Into Plugin Execution

MU mode goes further and reaches moment two. Perfmatters describes it as being able to “control every aspect of a plugin, from its front-end scripts, inline code, and MySQL queries (disable plugins entirely).” That is no longer asset management. That is deciding which plugins WordPress boots on a given URL, which is a different and considerably more powerful thing, and also a considerably easier thing to break a site with.

A caching plugin cannot do either of these things, and the reason is structural rather than a gap in anyone’s roadmap. A page cache’s job is to store whatever WordPress produced and hand it back unchanged. Deciding that WordPress should have produced something different is a job for code that runs inside the request, before the output exists.

Who Else Unloads Assets Per Page, and What They Charge

Perfmatters is the best known tool in this category. It is not the only one, and it is the only one here that costs money. All three directory figures below were read from the wordpress.org Plugin API on 25 October 2026. Note which products are absent: no caching plugin is in this table, because none of them does this job.

Asset CleanUp’s readme draws the same category line Perfmatters’ docs do, in one sentence: “Asset CleanUp is not a page-caching plugin. It controls what WordPress outputs; a page cache, server cache, or CDN can then store and deliver the optimized result.” Two independent vendors describing the boundary the same way is about as close to a settled answer as this field gets.

The free options are not free of cost, and Plugin Organizer says so on its own listing in capital letters: “WARNING: Reordering or disabling plugins can have catastrophic affects on your site.” Turning plugins off selectively is the highest-risk performance work available to a site owner, and the tools that make it easiest to do also make it easiest to do badly. Perfmatters’ Testing Mode, which applies your configuration only to logged-in administrators until you are satisfied, is a real and underrated part of what the $29.95 buys.

Neither is Perfmatters merely a script manager with extras. Its price also covers a code snippet manager with a conditions builder, local hosting for Google Analytics and Google Fonts, a login URL change, and speculative loading on hover. Whether those are worth paying for depends entirely on how many of them you already have.

Which Cache You Pair It With Decides Whether the Pairing Is Supported

Everything above treats “a caching plugin” as one category, and buyers do not. The three products weighed against each other most often here are WP Rocket, FlyingPress and Perfmatters: two caches and one script manager. So the decision is which cache the script manager sits beside, and the two premium caches answer that in opposite ways.

FlyingPress answers by refusing. Its one article on the subject names Perfmatters directly: FlyingPress “automatically disables conflicting optimizations” in Breeze, EWWW Image Optimizer, Perfmatters, ShortPixel Adaptive Images and SiteGround Optimizer. It closes with an instruction rather than a suggestion: “For best results, use FlyingPress as your only optimization plugin.”

Perfmatters answers by publishing instructions. Its documentation index holds 47 pages, and exactly two are settings guides for a named caching plugin: LiteSpeed Cache and WP Fastest Cache. Each says which of that plugin’s features to hand over. No WP Rocket or FlyingPress guide is in the index, and the obvious URLs for both returned 404. Read 25 October 2026.

The two positions are closer than they look. Both reach the same configuration, where one tool owns front-end optimization and the other stays out of it. Perfmatters documents that route by hand for two caches; FlyingPress walks it for you and tells you not to buy the second tool. The disagreement is about who holds the switch.

Neither vendor documents what a Perfmatters owner on FlyingPress most needs to know. FlyingPress disables conflicting optimizations without saying which Perfmatters settings those are. The Script Manager is the one feature FlyingPress has no equivalent for, so on a plain reading it should survive, which is not a basis for asserting that it does. Turn Testing Mode on, unload one asset, read the HTML.

FlyingPress Has No Script Manager Either, and We Counted

That cross in the table is a count, not an impression. Every article FlyingPress lists under Configuration and Features was fetched on 25 October 2026. There are 33, covering caching, assets, images, fonts and the database, and not one documents unloading a plugin’s stylesheet or script on a single URL.

The nearest thing it ships is Bloat Settings, a fixed list of nine switches: block editor CSS, Dashicons, emoji, jQuery Migrate, XML-RPC, RSS feeds, oEmbeds, post revisions and the Heartbeat API. All nine act on WordPress’s own output, site-wide, and most caching plugins ship the same list, ours included.

What the count does not show. An article count is not a feature count, and an undocumented feature is not always absent, the caveat our FlyingPress and WP Rocket comparison carried when it counted documentation another way. What it establishes is that a reader buying FlyingPress to drop Perfmatters finds the Script Manager’s job nowhere in its documented feature set.

The Per-URL Asset Census: Decide Before You Buy

Here is a method for answering the question from your own site rather than from anyone’s feature grid. It takes about ten minutes and needs no plugin, no account and no purchase. We have not seen it published anywhere, so treat the name as ours and the technique as yours.

Step 1. Turn combination off first. If your caching plugin is combining CSS and JavaScript into bundles, the census cannot see which plugin owns what, because the combine flattened the ownership away. Switch combine off, clear the cache, and measure before switching it back on. Skipping this step produces a reassuring zero that means nothing.

Step 2. Choose three URLs that represent your site. The homepage, one ordinary post or page, and your single highest-value page: the product page for a shop, the contact page for a consultancy. Three is enough to see a pattern and few enough to finish.

Step 3. List the plugin assets on each URL. One command per URL, from any terminal:

curl -s https://example.com/ | grep -oE '/wp-content/plugins/[a-z0-9-]+/[^"]+\.(css|js)' | sort -u

Step 4. Group by the folder name and mark each one. The folder immediately after plugins/ is the plugin slug. For each slug, answer one question: does this URL use that plugin’s feature? A form plugin on a page with no form is a no. An analytics plugin is usually a yes everywhere.

Step 5. Count the noes. That number is your unload budget: what a script manager can remove from this URL and what no page cache ever will, because a cache stores the page faithfully, including the parts nobody needed.

If the count is zero or one on all three URLs, a script manager will not repay a licence and your remaining time is elsewhere. If it is five or more on your most important page, you have found something real, and Asset CleanUp acts on it free before you decide whether Perfmatters’ wider toolkit earns the annual fee.

The Limitations of This Census

Stating these plainly matters more than the method itself, because the census is easy to over-read.

• 📏 It counts files, not bytes or execution time. Ten small stylesheets can matter far less than one large script that blocks rendering.
• 🙈 It only sees assets that appear in the HTML with a URL. Inline scripts, assets injected later by JavaScript, and anything served from a bundle are invisible to it.
• 🧩 A site that already combines assets scores near zero. That is an artifact of the combine step, not evidence of a lean site, which is why step one exists.
• 🤷 It cannot tell you whether an asset is needed. Only the person who built the page knows that, and getting it wrong breaks things.
• 🕐 It reads one moment in time. A plugin may enqueue conditionally, so a URL that looks clean today can load differently for a logged-in visitor or on mobile.

What the Census Found on Twelve Public WordPress Sites

We ran the census against twelve public sites on 10 September 2026 to show the spread, not to grade anyone. Eleven were detectably WordPress. Every number below is reproducible with the command above, and a high count is not a fault: it usually means the site does more things.

Three readings, and the third is the one that matters most. First, the two axes come apart: wptavern.com loads assets from only four plugins but 45 separate files, and 34 of those 45 come from a single plugin, the Gutenberg block editor. Counting plugins and counting files answer different questions, and a site can look busy on one axis and lean on the other.

Second, wpbeginner.com sits at the top of the table and its own asset list includes Perfmatters. A site that runs the tool still enqueues front-end assets from eleven plugins on its homepage, which is the honest shape of this work. A script manager does not empty the page. It removes what a URL does not need, and a large busy homepage needs a lot.

Third, the three zeros are the finding. Those sites are not loading nothing. They have combined their assets, so the plugin folder names no longer appear in their markup and the census has nothing to read. A zero on your own site is most likely that, and step one is how you get a real answer.

Running This Across Twenty-Five Client Sites

Everything above assumes one site and one person deciding. The arithmetic changes when the same decision has to be made twenty-five times, because per-page unloading is per-page work. A configuration built for one client’s homepage tells you nothing about another client’s, and the tool that makes each decision faster does not reduce the number of decisions.

We build xSpeed Cache, which is a caching plugin, so we sit on the left-hand column of every table in this article. Our free tier covers moments one and four: static file caching that answers before PHP starts, object caching, browser cache headers, preloading, minify, defer, delay, lazy load and database cleanup. Free versus Pro lists which side of the line each capability falls on, server compatibility covers what changes on nginx, and browser cache settings covers the headers. Pricing starts at $29/yr for Personal with a 14-day money-back guarantee, and every tier is also sold as a one-time lifetime licence.

Where fleets are concerned, the part we would point an agency at is not the cache. It is xSpeed Hub and the MCP server, which is free in our plugin where WP Rocket puts it behind a paid tier. One connection reaches every site you manage, so an assistant can read cache status, purge, run a scan and adjust settings across the fleet without opening twenty-five dashboards. On WooCommerce stores, checkout is never cached and logged-in customers never get a cached page.

Where Our Own Argument Is Weak

Four counts against us, and the first one is the whole subject of this article.

• 🚫 We do not ship per-page conditional unloading. Neither does any of the eight competitors in our comparison matrix, LiteSpeed Cache and WP Rocket included, and the count above puts FlyingPress in the same column. If that is the capability you came here for, we do not have it and Perfmatters or Asset CleanUp does. Nothing in our free tier or our Pro tier changes that answer.
• 📉 We are small. 10,000+ active installs and a 5.0 rating from 21 ratings, read from the Plugin API on 25 October 2026, against LiteSpeed Cache’s 7,000,000+ on the same query and Perfmatters’ claim of 185,000+ websites. That figure is a directory bucket, not a measurement, and we are still too new for the field datasets third-party roundups rank on.
• 🛠️ Perfmatters’ Testing Mode has no equivalent on our side. Previewing a risky configuration as an administrator before it reaches visitors is good engineering and we do not offer it for asset changes.
• 👥 Their support is the two developers who wrote the plugin. For a single site owner making a sensitive change, that is a real advantage over any larger team, and it is worth money.

Common Mistakes People Make Pairing These Two Tools

• ⚠️ Running both minifiers. Enabling minification in the caching plugin and in Perfmatters is the failure their own documentation warns about. Pick one implementation per feature and turn the other off completely.
• ⚠️ Running the census with combination on. The bundle hides the plugin folder names, so you get a zero that reads like good news. Census first, combine second.
• ⚠️ Unloading everywhere with no exception. Disabling a form plugin site-wide and forgetting the exception for the one page with the form is the single most common way this goes wrong, and it fails silently because the page still renders.
• ⚠️ Expecting a script manager to fix a slow server. Unloading assets does not move your response time. If your problem is a 900ms server response, the fix is caching, a better host, or both, and the number that actually moved across 413 sites says which of the two to reach for first.
• ⚠️ Buying an optimizer before confirming the cache works. More sites than you would expect have a caching plugin installed and not caching. There is a four-check method for proving it, and it takes two minutes.
• ⚠️ Using MU mode on a site you cannot roll back. Disabling plugin execution is the deepest change in this category. Take a backup you have actually restored once, and change one thing at a time.

Frequently Asked Questions

I bought Perfmatters and my PageSpeed score did not move. What went wrong?

Most likely you enabled features you already had. If your caching plugin was already minifying, deferring and lazy loading, adding a second implementation of the same three things changes nothing. The part of Perfmatters with no equivalent is the Script Manager, and it does nothing until you open it and unload something. Run the census in this article and see whether there is anything to unload.

Do I still need a caching plugin if I have Perfmatters?

Yes. Perfmatters does not do page caching, object caching, browser cache headers or cache preloading, and its documentation does not claim otherwise. It is designed to sit alongside a cache, not to replace one.

I turned on Perfmatters’ minify and my site broke. Why?

Check whether your caching plugin is also minifying. Two tools rewriting the same JavaScript is the classic cause, and Perfmatters’ own guidance is not to enable the same feature in both plugins. Turn minification off in one of them, clear both caches, and test again.

Is Perfmatters in the WordPress plugin directory?

No. A query to the wordpress.org Plugin API for the slug on 25 October 2026 returns Plugin not found, and so do WP Rocket and FlyingPress. There is therefore no public active-install count for any of the three, and any number you see quoted is the vendor’s own. Perfmatters’ site says 185,000+ websites.

I disabled a plugin’s scripts everywhere and my contact form stopped working, and my page builder broke too. How do I fix it?

Reopen the Script Manager, find the plugin you disabled, and add an exception for the URL or post type that needs it. Perfmatters supports exceptions layered on top of an everywhere-disable for exactly this case. It is the intended workflow rather than a workaround.

Can I get per-page script unloading for free?

Yes. Asset CleanUp Lite has 100,000+ active installs and a 4.9 rating from 1,564 ratings, and it unloads assets per page. Plugin Organizer and Freesoul Deactivate Plugins go further and disable whole plugins per URL. None of them includes the asset optimizer Perfmatters bundles alongside, so the comparison is capability against capability, not product against product.

My census shows zero plugin files. Is my site already perfect?

Probably not. Zero almost always means your assets are being combined into bundles, so the plugin folder names no longer appear in your HTML. Three of the eleven WordPress sites we checked returned zero for this reason. Turn combination off, clear the cache, and run it again.

Does the Script Manager affect the WordPress admin?

No. Perfmatters’ documentation states plainly that it works on the front end of the site. If your admin is slow, look at the Heartbeat API, object caching and your database rather than at front-end assets.

We manage 25 client sites. Does per-page unloading scale?

The tool scales. The decisions do not. Each site needs its own configuration because each site runs different plugins on different templates, and no export from one client helps another. Budget the time per site rather than per licence, and start with the two or three sites where the census count is highest.

I bought Perfmatters and then moved the site to FlyingPress. Do the two work together?

Not as a supported pairing. FlyingPress names Perfmatters as one of five plugins whose conflicting optimizations it disables automatically, and tells you to run FlyingPress alone. The section above has the quotes and the caveat. Prove your own case with Testing Mode before you renew either licence.

How much does Perfmatters cost, and how does that sit beside the caches?

$29.95 a year for one site, $59.95 for three, and $124.95 for unlimited sites with multisite support, re-read from its pricing page on 25 October 2026. Every tier carries a 30-day money-back guarantee, renewals carry an automatic 15% discount, and the page currently shows a public code for a further 10% off. The pairing table above puts that beside each cache’s own licence.

Does xSpeed Cache work alongside Perfmatters or Asset CleanUp?

Yes, and the same rule applies as with any other pairing. Let one tool own each feature. If you want Perfmatters handling front-end assets, use our caching, object cache and preloading and turn our overlapping optimization modules off. Our migration module imports settings from competing cache plugins, which is a different job from asset management and does not conflict with it.

Conclusion: Which One Do You Actually Need?

The honest answer for most sites is that these are not alternatives. One acts before PHP runs and the other acts while PHP is running, and a site with a real problem at both moments needs both. A site with a problem at only one of them is paying twice if it buys both.

What to do this week:

1. Prove your cache is running before you change anything else. Two commands and four checks settle it.
2. Run the per-URL asset census on three pages, with combination switched off first. Write the numbers down.
3. If the count is high, install Asset CleanUp Lite free and unload one thing. If the workflow suits you and you want the wider toolkit, Perfmatters at $29.95/yr is the mature product in this space and we would not argue against it.
4. If the count is low, your remaining time is in caching and hosting instead. Run the free xSpeed Scan on the URL, which needs no account, and read the server response number first.
5. If your server response is the large number, a faster origin is the fix. We recommend xCloud, and it is ours: xCloud and xSpeed Cache are both Startise companies, and it gives a cache an origin tuned for it.
6. If you manage more than five sites, our free tier and the fleet tools are where we would start you, and the 80-capability comparison shows which tier every feature ships in.

We build a caching plugin, so we sat out the one row this article is really about, and we said so rather than blurring it. That is deliberate. These tools are not competing for the same job, and a reader who knows which moment their problem lives at spends less and gets more. What we build is the other half of that answer: a cache and an optimizer for people running more than one WordPress site who want the whole fleet reachable from one connection, on a free tier that does not expire and never becomes a trial. Run the free scan on your slowest URL and start with whichever number comes back largest.