Google Search Appliance (Google mini) for startups

Question

Is Google Search Appliance a good fit for a startup like Stackoverflow (just an example)? How does it fare?

I haven’t seen many examples or talk about this device being used in the startup circles. I need enhanced search experience with capabilities like stemming, spell check, relevancy, easy of indexing and retrieving and complete customization of the user interface (more than what Solr-Lucene or Autonomy provides today. In fact I am bowled over by Google Search Appliance labs). Also speed is of prime concern. With this in perspective I wonder why GSA is not gaining adoption among the non-enterprise crowd. Is there any reason in particular?

Additionally, do they have flexible pricing models for startups ($30,000 for 500,000 documents seems exorbitant at this time)? This is to me is the primary deterrent...

Answer 1

I'm not sure if you are asking about intranet or website search, but in either case enterprise products are designed for enterprise problems, which tend to be very different from startup problems.

Large companies tend to have large numbers of unstructured documents, built up over years, that could be stored anywhere in any format, with the author probably having left the company. On the website side, they have things like support documents for long obsolete products. The only practical way to organize all this is a general purpose search engine, like the one Google uses to index the web.

On the intranet side, startups tend to have a relatively small number of documents stored consistently in one place, probably in something like Google Docs which has good search functionality already.

Startup websites are almost by definition going to be doing something different from everyone else, which doesn't make a good market for a one size fits all search product. They are likely to be either dealing with some sort of structured data (so a general purpose search engine is either unnecessary or inadequate) or need something specialized that no existing product will handle.