🛠️Probabilistic Matching
Also known as Probabilistic Identity Matching, Fuzzy Matching, Probabilistic ID
A best guess that two records are one person.
Think of it like
It is like guessing who wrote a note from the handwriting. You are often right, not always.
Example
A phone and a laptop use the same home Wi-Fi every evening. A tool guesses they belong to one person and joins them, with a score of 0.8.
Why it matters
It fills the gaps that exact IDs leave. But each wrong guess sends the wrong ad or bends a report.
Definition
Common mistake
Key takeaways
- 1Clues, not IDs. Shared signals stand in for a login.
- 2A score, not a fact. Each join is a guess with odds.
- 3Privacy limits it. Browsers hide more clues each year.