Recent Tweets
join our mailing list
* indicates required

About Ambleside

Ambleside Logic is led by Aaron Rosenbaum. Father of 3, Programming since 7, DevOps since 11 (hacking RSTS), exIngres, exCTP, exCohera. Sold two companies to Oracle, one to HP. Research + Strategy for NoSQL/BigData ecosystem implementors, vendors and investors.

« Virtualization for SMB, an Oxymoron? | Main | Hulu Plus + Pandora on Tivo - "We have a pulse!" »
Thursday
Apr072011

Why does product search still suck?

In the past few years I've been away from enterprise and commerce search, it seems like things have stood still.

Hadoop, Lucene, Solr, Nutch - all very cool.  But where are the taxonomy tools? Has anyone improved on the Potters Wheel update we did at Cohera?

Why do I say this? Lets pick on one site:

Run "Kids Hockey Movie" on Netflix.com

It doesn't return "Mighty Ducks" on the first page.  Ack!

Take a look at slots 4+5 - Scary Movie and Garbage Pail Kids!

Why? Seems like a weak synonym library or a broken content process.  Or can Solr not handle this? I doubt it.

Looking at the content, the description of "The Mighty Ducks" doesn't contain the common dictionary synonyms for kids.  They are a "peewee" hockey team.  The merriam webster doesn't show peewee as a synonym for kids nor is the inverse listed.

That doesn't excuse this.

1) If you are going to have a real meta-tagging process, which netflix does, then you should build special keyword libraries..."peewee" needs to find sports stuff unless "herman" is in the query too.  

2) Analytics tools should be able to rank mis-searchs and throw resources at the special cases - especially with such a limited content body as Netflix.

"Hockey Peewee" does find it.

What other sites have search that still sucks?

PrintView Printer Friendly Version