Not quite peerless,
but it could get there

Peer39 released SemanticMatch, which matches ads to content by processing natural language, meaning and sentiment on websites.

"We eliminate the errors that can plague keyword targeting," said CEO Amiad Solomon. "And unlike other forms of online targeting, Peer39’s SemanticMatch does not set cookies or track user behavior."

Advertisers may terrace users with criteria like Awareness ("autos," for example), Consideration ("SUV"), Preference ("hybrid"), Purchase and Retention. The system does not require keyword targeting.

Matthew S. Goldstein recently left his post as SVP of Revenue Operations at TACODA to join Peer39 as COO. TACODA is one of AOL's major behavioral targeting arms.

"When I studied the next generation of ad targeting technology, it was clear immediately that Peer39 has taken granular targeting to its deepest level by understanding page meaning and sentiment," Goldstein declared.

Peer39 launched early last year. It drew $3 million in Series A funding and another $8.2 million six months later.

A number of other companies have tried building viable algorithms for "natural language" search and contextual advertising, including Powerset, which is currently testing its capabilities on Wikipedia.

The most notable natural language search attempt — and failure — was the Ask Jeeves search engine. Having ditched both Jeeves and natural search long ago, Ask recently narrowed its focus to the niche it (hopefully) knows best: married women.