Why is simply removing the identifying attributes (like names) not sufficient for anonymization?

Because quasi-identifiers remain — and combinations of them can still single out individuals.

Removing the name feels decisive, but the quasi-identifiers are still sitting there untouched. Knowing someone is a 32-year-old male with a beer rating of exactly 8 can be enough to identify them in a small or known group. The record is "nameless" but not anonymous.

This is the core trap of de-identification: identifying attributes are the obvious risk, quasi-identifiers are the real risk. Effective anonymization must process the quasi-identifiers — generalizing or suppressing them — so that each person blends into a crowd of similar records.

Tip: Name removal is step zero, not the finish line. If you stop there, you've built a false sense of safety.

Go deeper:

• k-Anonymity (Sweeney, 2002) — shows quasi-identifiers alone re-identify people.
• Data re-identification (Wikipedia) — name removal is provably insufficient.