IP-based tracking is one of the oldest tracking mechanisms on the web — and also one of the least precise on its own, worth understanding both how it works and why it rarely identifies a specific individual by itself.
The basic mechanism
Every request to a server includes the requester's IP address, which servers commonly log. Repeated visits from the same address can be correlated across sessions, forming a basic (if crude) tracking mechanism that requires no cookies or scripts.
Why shared addresses complicate this
Because home networks, mobile carrier gateways, and corporate networks frequently share one public address across many actual users (see Public IP vs Private IP and Mobile IP Addresses and Carrier-Grade NAT), "the same IP visited twice" doesn't reliably mean "the same person visited twice" — it might be two different household members, or two entirely different mobile customers sharing a carrier gateway.
Why address changes complicate this further
Dynamic IP addresses change over time (see Static IP vs Dynamic IP), meaning IP-based tracking alone loses continuity whenever an address rotates — a real limitation for anyone relying on IP alone as a long-term identifier.
Why an IP alone rarely identifies one person
Combining the shared-address and address-rotation issues, IP-based tracking is a weak signal in isolation. It becomes far more powerful when correlated with other signals — cookies, account logins, browser fingerprinting — which is exactly why modern tracking rarely relies on IP address alone, treating it as one input among several rather than a standalone identifier.
FAQ
Can changing my IP address alone stop tracking?
Not by itself if the tracker is also using cookies, account logins, or fingerprinting — those other signals persist across an IP change.