What Does “Clearing Cache” Actually Clear?

The advice everyone gives and nobody explains

I have lost count of how many times, during years of supporting websites and teams, I have typed the sentence “try clearing your cache.” It works often enough that it has become a ritual, like switching a device off and on again. But rituals without understanding cause trouble. People clear everything, lose their saved logins, and then complain that the fix “broke” something. Others never clear anything and wonder why a website still looks like last year’s version.

The truth is that “the cache” is not one thing. There are many caches, living in different places, owned by different people. When someone says clear it, they might mean your browser, your phone app, your computer’s address book, a content delivery network, or a plugin on a web server. Knowing which one is the difference between a ten-second fix and an hour of confusion.

A cache in everyday life

Before the computer version, think of a kitchen. You could walk to the shop every time you need salt. Instead, you keep a small jar on the counter. Fetching from the jar takes a second, going to the shop takes twenty minutes. The jar is a cache.

It has the same properties as every digital cache. It is small, so it can only hold what you use often. It is a copy, so the shop still has the real thing. And it can go stale: if the shop changes the recipe on the salt, your jar still holds the old version until you refill it.

That last point is the source of nearly all cache problems. A cache is fast because it trusts old copies, and it is annoying when the trust is misplaced.

Why caches exist: speed, in numbers

Computers have a hierarchy of storage. Closer to the processor means faster and smaller, further away means slower and bigger. As a rough scale (these are order-of-magnitude figures that vary by machine), reading from the CPU’s own cache takes about a nanosecond, from main memory about a hundred nanoseconds, from a solid-state drive around a hundred microseconds, and from a server across the internet tens of milliseconds. That is a gap of millions of times between the fastest and slowest layers.

Caching exploits that gap. The system keeps copies of frequently used things closer, so it can skip the slow trip. A simple calculation shows how much this matters. Suppose a cache hit takes 5 milliseconds and a miss, which means going to the original source, takes 200 milliseconds.

With no cache at all, every request waits 200 ms. If 85% of requests are hits, the average wait drops to 0.85 × 5 + 0.15 × 200 = 34.25 ms, nearly six times faster. At a 95% hit rate it is under 15 ms. The curve is a straight line, so every additional percent of hits is worth the same amount. That is why companies spend real money on caching.

The layers: who is caching what

This is the part most guides skip. Between you and a website, copies are stored in at least five different places.

• CPU cache: tiny, ultra-fast memory built into the processor. It manages itself and you cannot clear it.
• Browser cache: your browser saves images, stylesheets, scripts and fonts so repeat visits are quick. This is what “clear cache” normally refers to.
• Operating system and DNS cache: your computer remembers which numeric address belongs to a website name so it does not ask again every time.
• CDN cache: a content delivery network keeps copies of site files on servers around the world, close to visitors.
• Server cache: the website itself may save ready-made versions of pages, so it does not rebuild them for each visitor.

A reader can clear only the second and third. The last two belong to whoever runs the site. If you publish websites yourself, as many of my readers do, you will deal with all five.

What your browser cache actually contains

Open any web page and your browser downloads a bundle of files: the page itself (HTML), stylesheets (CSS) that control appearance, scripts (JavaScript) that add behaviour, images, and fonts. Many of these rarely change. The logo on a site is the same on every page, so downloading it fresh each time would be wasteful.

The browser therefore stores these files, and the next time it needs them it can skip the download. Here is what that looks like for a typical page, using realistic but illustrative sizes.

On a first visit this example page downloads 1,850 KB. On a repeat visit, only the 60 KB HTML needs to be fetched, a saving of about 97%. At a typical home connection of 10 Mbps, the first load needs about 1.5 seconds of transfer while the repeat takes a fraction of that. On a slow 2 Mbps mobile connection, the first visit would need around 7.4 seconds of transfer alone, which is exactly why the cache matters so much on phones.

How the browser decides whether to trust its copy

A cache that never refreshed would show you stale pages forever. So websites send instructions, in hidden headers, telling the browser how long a file stays fresh. A typical header says something like Cache-Control: max-age=86400, which means “this file is good for 24 hours.” Within that window, the browser uses its copy without even asking.

When a file goes stale, the browser does not always download it again. It can ask the server a cheaper question: “I have the version labelled with this tag. Has it changed?” The label is called an ETag. If the answer is no, the server replies with a tiny “304 Not Modified” message and the browser keeps using its copy. Only if the file changed does the server send the full new version.

Developers also use a trick called cache busting. When they update a stylesheet, they change its name or add a version number, for example style.css?v=42, so that the browser sees a brand-new file and fetches it. When this is forgotten, you get the classic symptom: new page, old styling.

Cache, cookies and history: the three things people confuse

This is where most of the damage and most of the fear come from. The “clear browsing data” screen in a browser lists several separate items. They are different things.

The practical rule: if you only tick “cached images and files,” you will not lose logins or passwords. The mistakes happen when people press the big “clear all” button and tick every box. A moment of care is worth saving yourself a password-reset afternoon.

Five situations, five different fixes

Tap through the tabs. Each one shows a real problem and which layer to clear.

Windows / Linux:  Ctrl + Shift + R   (or Ctrl + F5)
macOS:            Cmd + Shift + R

Android: Settings > Apps > [app] > Storage > Clear cache
# Clear data / Clear storage also resets the app and signs you out

Chrome: lock icon > Site settings > Delete data
# or Settings > Privacy > Delete browsing data > Cookies

# Windows
ipconfig /flushdns

# macOS
sudo dscacheutil -flushcache
sudo killall -HUP mDNSResponder

1. Save or update the post
2. Purge the page cache in your caching plugin
3. Purge the CDN cache (for example Cloudflare: Caching > Purge)
4. Hard refresh your browser
5. Test in a private window

Notice that “clear the cache” was the right answer in only two of those five situations. In the others, the fix involved cookies, DNS, or the layers a site owner controls. That is the lesson of this whole article: ask which cache before you clear.

Hard refresh versus clearing everything

If a single website looks wrong, you rarely need to clear your entire browser cache. A hard refresh bypasses the cached copies for that page. On Windows and Linux you press Ctrl+Shift+R (or Ctrl+F5), and on a Mac, Cmd+Shift+R. It is the scalpel; clearing the whole cache is the sledgehammer.

Another trick I use constantly is the private or incognito window. It starts with no stored cache or cookies for that session, so if the site looks right there, you know the problem is a stale copy in your regular browser. If it looks wrong there too, the problem is not your cache at all: it is the server, the CDN, or the site itself.

If you run a website: the layers pile up

Readers have one cache to worry about. Site owners have several, and that is why “I updated it but it did not change” is such a common complaint. Imagine you fix a typo in a WordPress post and click Update.

• The page cache plugin may keep serving the pre-built old page until it is purged.
• The CDN may keep serving its copy from servers near your visitors until its rules expire or you purge it.
• Each visitor’s browser may keep its own copy until it is stale.

The order matters. Clear from the source outward: plugin first, then CDN, then your own browser. If you clear the browser first, it simply re-downloads the stale copy from the layer above it. And always verify in a private window, because your own normal window is the layer most likely to mislead you.

The same thinking applies to design changes. Many themes and plugins also bundle CSS and JavaScript files. If a style tweak does not show, purge the plugin’s file cache, and then purge the CDN. A teammate of mine once spent a whole morning editing a stylesheet that was working perfectly; the browser was just refusing to fetch it.

Mobile phones: cache versus data

Apps cache aggressively. A social app saves thumbnails and video chunks so scrolling feels smooth, and over months this can grow into gigabytes. On Android, you can open the app’s storage screen and choose to clear the cache, which frees space and costs you nothing but a slightly slower first scroll afterward.

The same screen has a second button, usually called “Clear data” or “Clear storage.” This one is different and more drastic: it resets the app, signing you out and removing settings and sometimes locally stored files. People hit the wrong button all the time. Read the label twice.

On iPhone there is no per-app cache button. Options are to offload the app, which keeps its documents but removes the app files, or to delete and reinstall it. Safari has its own setting to clear history and website data, which removes the browsing history, cookies and cached files together.

When clearing the cache is a bad idea

• Doing it daily out of habit. You throw away the speed benefits and re-download everything again and again.
• Doing it on a metered connection. Re-downloading a heavy site’s files uses your data allowance.
• Believing it protects your privacy. Cache is a small part of what sites know about you. Cookies, accounts and network logs are separate.
• Using it as a cure for everything. Slow internet, a failing disk and an overloaded server will not improve.

Six myths about clearing cache (tap to open)

Quick quiz: test yourself

Tap each question to reveal the answer and the reasoning.

Frequently asked questions