← All articles

How Do I Know If My Fingerprint Is Unique?

2026-08-25 · 6 min read

Test pages will tell you, and a high uniqueness score is bad news rather than good. Fingerprinting identifies you by rarity: a browser sharing its exact signal set with millions of others is effectively anonymous, while one that is one-in-a-million is trackable across visits without a single cookie. The goal is to be ordinary, not distinctive.

What does a uniqueness score actually measure?

A test page collects your signals, compares them against its own database of previous visitors, and reports how many of them matched. A result of "1 in 3 browsers share your fingerprint" means you are in a large crowd. A result of "you are unique among 500,000 tested" means the site could recognise you on any future visit, from any account, with no stored identifier.

The number is relative to that page's visitor pool, not to the whole internet. Different testers give different results for the same browser, which is expected and not a malfunction.

Which signals contribute the most rarity?

Entropy is not evenly distributed. A handful of fields carry most of the identifying weight, and they are the ones tied to hardware and installed software rather than to preferences. Changing the low-entropy fields does almost nothing; the high-entropy ones decide the outcome.

Signal	Entropy	Why
Font list	Very high	Reflects every installed application
Canvas / WebGL hash	Very high	Depends on GPU, driver and OS together
GPU renderer string	High	Names an exact chip model
Screen size and pixel ratio	Medium	Common sizes are shared widely
Timezone	Medium	Roughly 40 in common use
Language	Low	A few dozen values dominate
Browser version	Low	Everyone updates on the same cycle

A designer's machine with 300 installed fonts is more identifiable than a fresh laptop with 60, before anything else is considered.

Why does trying to hide make it worse?

Most attempts at hiding increase rarity. Blocking canvas entirely puts you in the small group of browsers that refuse to render, which is itself distinctive. Reporting 128 CPU cores to defeat hardware profiling produces a value no consumer device has. Randomising each field independently generates combinations that correspond to no real machine.

The pattern is consistent: any change that makes your browser unusual raises the score you were trying to lower. Blending requires looking like a common device, not like a defended one.

How do you read your own result?

Run two or three different test pages and look for agreement rather than a single number. Check the fields each page flags as your rarest, because that is where your entropy is concentrated. Then reload each page and confirm the values stayed identical — instability is a separate and worse problem than rarity.

Result	What it means	Action
Shared with many, stable	Ideal	Leave it alone
Rare but stable	Trackable across visits	Move toward common values
Common but unstable	Looks manipulated	Stop randomising
Rare and unstable	Worst case	Rebuild the profile

How much does a unique fingerprint actually cost you?

Uniqueness only matters where it is being used. For anonymous browsing it means an advertising network can follow you between sites without a cookie. For a logged-in account it means the platform can link that account to every other account opened from the same machine, which is the more expensive of the 2 consequences.

Sellers running 5 or 10 storefronts feel the second version first. One flagged account leads to a device-level check, and a device shared by 10 accounts turns 1 problem into 10 within a day or 2.

What does a good target look like?

Aim for a profile that resembles a mainstream laptop in the country its IP belongs to: a common screen size, a default font set, a GPU model that sells in volume, and a timezone matching the address. That combination is shared by a large population, which is exactly the point.

Check the score once when the profile is created and again after any browser update, then leave it alone. Chasing a lower number by adding more countermeasures usually moves it in the wrong direction, because each countermeasure is itself a signal that most visitors do not carry.