Table of Contents
Key Takeaway
- 🖐 The idea: Browser fingerprinting identifies you by the technical features your browser reports, including screen size, fonts, GPU, time zone, and language, stitched into a profile that survives cookie deletion.
- 📊 The scale: Research measured device fingerprinting accuracy at 90 to 99 percent as early as 2017, and a user in the Electronics Frontier Foundation’s own sample found a fingerprint unique among more than 241,000 test runs in 45 days.
- 🚫 The trap: The privacy habits people trust most, private windows, deleting cookies, and VPNs alone, do not stop browser fingerprinting, because the fingerprint lives in your device’s configuration, not in stored files.
- 🛡 The upside: The same technique is what banks and e-wallets use to flag someone else logging in as you, so the goal is controlled exposure, not silence.
- 🧭 Today’s move: Test your own browser on EFF’s Cover Your Tracks, then pick a browser whose anti-fingerprinting style fits your threat model before your next login-heavy session.
You deleted the cookies, you opened a private window, you even connected the VPN, and the travel website still greets your laptop like an old friend. That is browser fingerprinting at work: a tracking method that does not need to store anything on your device at all, because you are the identifier. Your screen resolution, your installed fonts, your graphics card’s quirks, your time zone, and dozens of other technical details combine into a configuration that research has shown can point to a single machine with 90 to 99 percent accuracy. Privacy tools that rely on deleting stored data never touch it.
This article explains what browser fingerprinting actually collects, why the countermeasures people rely on fail, what genuinely reduces it, and the trade-offs each choice carries. Every claim traces to a primary source or a named study, and where the evidence is a single tool’s sample, the article says so.
The browser fingerprinting numbers, verified one line at a time
| Item | Verified figure | Source |
|---|---|---|
| Identification accuracy | Device fingerprinting can identify users with 90 to 99 percent accuracy, even across multiple browsers, per research reported in 2017 | Ars Technica report on academic research |
| Entropy example | One Cover Your Tracks sample run estimated a browser’s fingerprint at “at least 18.52 bits of identifying information” | EFF Cover Your Tracks sample result |
| Uniqueness example | A test run reported a browser fingerprint “unique among the 241,379 tested in the past 45 days” | EFF Cover Your Tracks sample output, via user reports |
| Research dataset | AmIUnique, the academic fingerprinting study, reports identifying users among a dataset of more than two million collected fingerprints | AmIUnique project |
| Attribute count | Fingerprint collections commonly use dozens of attributes, with write-ups of one research dataset listing more than 75 attributes per fingerprint | AmIUnique dataset write-up |
| Browser defenses | Firefox ships built-in fingerprinting protection that reduces how much identifying information its browser leaks | Mozilla’s fingerprinting documentation |
One honesty note sits under the table: the two “unique among” examples are single-sample observations from public test tools, not population statistics. The instruments are real and the direction is unambiguous, but the exact number in your own test will be your own.
What Browser Fingerprinting Is, and How It Differs From Cookies
A cookie is a thing a website stores inside your browser, which is why you can clear it and why laws regulate it. Browser fingerprinting is the opposite: the site reads your browser out loud. When a page loads, scripts ask your browser standard technical questions, which screen size are you, which fonts do you carry, which graphics card renders you, what time zone are you sitting in, how does your audio stack behave, and each answer is a trait. Stack enough traits together and the combination belongs, in practice, to one machine rather than to a crowd.
The researchers behind the AmIUnique study describe the technique as reading device and browser configuration to build that profile, and Mozilla’s documentation lists the raw materials: browser type and operating system, screen resolution and device type, installed fonts, language and time zone settings, hardware and graphics details, the user agent string, the IP address, and the way the browser responds to specific scripts and APIs. None of these traits identifies you alone. Two people in Manila can share a laptop model, a language, and a time zone. The fingerprint works because collisions across fifteen or twenty traits at once are rare, and the rarer the combination, the stickier the identifier.
Two properties make this more durable than cookie tracking. First, there is no file to delete, because the identifier is recomputed from your settings every time you load a page. Wipe your cookies and the fingerprint returns identical. Second, it crosses sites and sessions without any cooperation between the sites, which is why the technique became the tracking layer for advertisers after browsers tightened their cookie rules.
What Your Fingerprint Is Made Of
The traits fall into four families, and knowing them tells you why each countermeasure works or fails.
Hardware and rendering: screen resolution and color depth, device pixel ratio, the graphics card and driver strings exposed through rendering APIs, processor class, device memory, and touch hardware. These are the deepest traits, and your browser exposes them for a good reason, because websites must render correctly on your screen. Audio-stack behavior and canvas rendering output sit here too: two identical laptops can render a hidden drawing, or process a tone, in subtly different ways.
Software and fonts: operating system and version, browser family and version, installed plugins and extensions, system fonts, and which media codecs your machine can play. The font list alone is famously identifying, because installed fonts quietly record every design app and language pack a machine has ever acquired.
Settings and locale: language preferences, time zone, whether cookies are enabled, do-not-track flags, and your user agent string. Time zone and language are the quiet workhorses of geolocation-by-configuration, and they survive every browser restart.
Behavior and network: typing rhythm in advanced setups, battery status in earlier mobile implementations, and your IP address. The IP is the trait a VPN can genuinely move, which is why vendors oversell it, and why the next section matters for anyone paying for one for privacy reasons.
Our own browser extension security audit covers a related leak family, extensions, because every extension you install changes the fingerprint and can open its own exposure. The same logic applies to fonts and plugins: every tool you add is another detail someone can collect.
Why Incognito, VPNs, and Deleting Cookies Do Not Stop It
Run the standard privacy checklist against browser fingerprinting and you get an uncomfortable score. A private window changes what gets stored, not what gets read, so the fingerprint returns the same. Deleting cookies removes stored identifiers and leaves the configuration untouched. A VPN replaces the IP address, one trait in a stack of dozens, while all the local traits keep reporting, which is why privacy researchers treat a VPN as a network tool and not an anti-fingerprinting tool. Even the Tor Project’s own user forum carries complaints from people who ran Tor and still saw their browser flagged as nearly unique, because in a small crowd every member is unusual.
There is also the anti-pattern to name: fighting fingerprinting by changing settings one by one can make you more unique, because a crowd that all uses defaults is anonymous by sameness. One person running an unusual screen size, a disabled API, and a nonstandard font set is the loudest fingerprint in the room. This is why serious defenses, discussed below, are built by browsers with large user bases and shipped to everyone at once, rather than assembled by hand.
For balance, the technique has a legitimate face that protects you. Banks, e-wallets, and licensing platforms use device and browser fingerprinting as a fraud signal: if a login claims to be you but arrives with a fingerprint that has never been seen on your account, asking for another verification step is exactly the right behavior. We covered the login side of the same pattern in our passkeys guide for Southeast Asian banks and wallets. The fight, honestly framed, is not to be invisible, which nothing achieves, but to stop being the only one of your kind in the market for ad-targeting data.
What Actually Slows It Down
Three real strategies exist, each with a cost, and the right one depends on what you are protecting.
Blend into a large crowd. Use a browser that millions of people use in its default configuration, on a common device, with default settings, in the most common time zone and language for your region. This is generalization: you are hard to identify because you are ordinary. The cost is real, because you give up the customizations that make a machine yours, and on a long commute the same sameness that hides you also disables your tools.
Break the reuse of the fingerprint. A browser that actively randomizes or partitions what scripts can read, such as Brave’s randomization of noise in readouts or Tor Browser’s strict uniformity, attacks the repeatability that makes fingerprinting valuable. The cost shows up as breakage and friction: sites that fail to render, video players that misdetect, and more robot-checks, because anti-bot systems read the same signals fraudsters hide behind.
Turn on the shipped protection and keep the crowd. Firefox ships fingerprinting protection designed to reduce how much identifying information leaks, and Chromium-family browsers now ship privacy controls in the same direction, but the protection only helps at population scale, which is the trade for staying usable. This is the choice most professionals should make about browser fingerprinting, because the goal is not to become anonymous, which is practically impossible, but to stop being uniquely trackable across every site on the internet.
What does not work is worth repeating once more in plain words: incognito windows, cookie wipes, and a lone VPN do not slow fingerprinting, and any product sold to you primarily with those claims deserves a second read.
What a Professional Should Do This Week
- See your own fingerprint. Browser fingerprinting is best met with evidence: run EFF’s Cover Your Tracks or AmIUnique once, read the “bits of identifying information” line and the “unique among” phrasing, and run it again in a different browser to feel the difference between them. The comparison teaches more than any article, including this one.
- Match the browser to the threat. For daily work where sites must function, a mainstream browser with fingerprinting protection on, plus the extensions discipline from our audit piece, is the durable answer. For research sessions where breaking sites costs nothing, Tor or a randomized private browsing profile fits. For banking and e-wallet logins, let fingerprinting work for you: it is the fraud signal that flags someone else logging in as you.
- Stop pretending the VPN is a privacy suit. Keep it for what it does, network privacy and geo-unblocking, and stop expecting it to hide your browser fingerprinting configuration, which it cannot. Network privacy is a different job, and our VPN setup guide covers that job honestly.
- Keep the trait list short. Fewer extensions, fewer fonts imported, and fewer special settings all shrink the fingerprint. Every tool you install is a trait you publish, as our extension audit showed with real permission data.
- Protect the logins the fingerprint guards. If a bank or wallet ever asks for verification because your device looked new, treat that as the system working, confirm it was you, and never bypass the extra step, because the request exists precisely because fingerprinting and fraud detection are doing their job.
Frequently Asked Questions About Browser Fingerprinting
Does incognito mode stop browser fingerprinting?
No. A private window changes what your browser stores, not what it reports, and browser fingerprinting works exactly the same inside it. The technical traits, screen, fonts, graphics behavior, time zone, are read live every session, so the fingerprint stays identical whether you browse privately or not.
Does a VPN stop browser fingerprinting?
Only the network portion. Browser fingerprinting reads configuration, and a VPN moves your IP address, one trait among dozens, while the local hardware, software, and settings keep reporting, so a VPN alone reduces nothing about fingerprint uniqueness.
Is browser fingerprinting legal?
In most jurisdictions it currently sits in a gray zone: it is not banned outright, but transparency and consent rules that cover online tracking apply to it unevenly across privacy laws, which is exactly why it grew after cookie rules tightened. Check your local data privacy regulator for the rules that apply to you.
How can I see my own fingerprint?
Run EFF’s Cover Your Tracks or AmIUnique. Both report how identifying your browser is, in bits of entropy and in “unique among” phrasing, and AmIUnique explains the research behind its dataset of millions of fingerprints.
Can browser fingerprinting be turned off?
Not completely, because the same browser traits that track you are the ones websites need to render correctly. The realistic options are blending into a large default crowd, using a browser that randomizes or partitions the readable traits, or enabling shipped protections, each with usability costs.
Is fingerprinting good or bad for my bank account?
Both, honestly. The same signals that let advertisers follow you let your bank and e-wallet tell when a login is not really you, which is why unexpected verification steps are a protection working, not a malfunction.
- coveryourtracks.eff.org — EFF Cover Your Tracks
External sources linked in this article, in order of first mention. Spotted an error or a broken link? Tell us.
Cite this pagePublished Nov 20, 2026
Agron, E. (2026, November 20). Browser Fingerprinting: the Tracking That Incognito Never Stops. WorldNgayon. https://worldngayon.com/browser-fingerprinting/
Agron, Edmon. “Browser Fingerprinting: the Tracking That Incognito Never Stops.” WorldNgayon, 20 Nov. 2026, worldngayon.com/browser-fingerprinting/.
Agron, Edmon. “Browser Fingerprinting: the Tracking That Incognito Never Stops.” WorldNgayon, November 20, 2026. https://worldngayon.com/browser-fingerprinting/.
Agron, E. (2026) ‘Browser Fingerprinting: the Tracking That Incognito Never Stops’, WorldNgayon, 20 November. Available at: https://worldngayon.com/browser-fingerprinting/.
E. Agron, “Browser Fingerprinting: the Tracking That Incognito Never Stops,” WorldNgayon, Nov. 20, 2026. [Online]. Available: https://worldngayon.com/browser-fingerprinting/
@online{agron2026browserfingerprinting,
author = {Agron, Edmon},
title = {{Browser Fingerprinting: the Tracking That Incognito Never Stops}},
year = {2026},
date = {2026-11-20},
organization = {WorldNgayon},
url = {https://worldngayon.com/browser-fingerprinting/},
urldate = {}
}TY - ELEC AU - Agron, Edmon TI - Browser Fingerprinting: the Tracking That Incognito Never Stops T2 - WorldNgayon PY - 2026 DA - 2026/11/20/ UR - https://worldngayon.com/browser-fingerprinting/ Y2 - ER -
Cite the version you read; the page may be updated. Zotero and Mendeley can also import this page directly.






