The New York Times technology blog, BITS, had a post, Google Goes Corporate, which talked about their recent acquisition of Postini, which “offers a service to help companies protect and control their e-mail , for $625 million in cash.” Now the title is a bit misleading as Google has been trying to go corporate for a number of years. I am sure that they will succeed on some level and making the right acquisitions will be key.

While I use Google all the time and would be lost without it, there are many limitations to the core search tool that may have greater impact in the enterprise than out on the web. Personal intent is one. For example, I did a post on “Groove on Location in Albania” in 2004 about the interesting work a friend was doing with Groove software in Albania. When I typed in - location Albania - in Google in 2004 I got 836,000 hits. My post was on the first page with many Albanian dating services, a listing of air conditioning contractors, and how to get phone cards. Now in 2007, my post is still number 8 but its companions are quite different but they still represent a variety of intentions.

Google is also optimized to look at the unstructured data found on the web but inside the enterprise there is also much structured data. There are many options to the reliance on text-matching links to unstructured data that Google offers. For example, EasyAsk offers the ability to put in plain language queries and have the search engine respond back with a clarifying question to narrow the search. For example, if you ask, “what is the price of a stock trade” on a financial services site Google and other text-matching tools will have many types of responses. EasyAsk might come back and ask you, “do you mean commission and fees?” which contains none of your original words.

Another feature found in other search tools is the concept of “termless” search. For example, iQuest which, like EasyAsk, can be targeted intelligently at both structured and unstructured data, also can let content speak for itself. The tool surfaces the top number of (usually 10 – but that number can be set) concepts or themes that are being discussed in a given body of text such as a blog. This reduces much manual analysis and allows you to look deeper into what people are saying in large amounts of content. It also looks at everything and is thus does not let bias creep. Most manual analysis by people will have preconceived notions that consciously or even unconsciously sift out what might seem important. The automated tool is not “smart” enough to do this which is a good thing.

These are just some of the features enterprise Google will need to work on if it is to compete on more than price and brand.