A couple weeks ago we reported a rumor that Google was going to replace its Mini search appliance with a new hosted solution for indexing content found behind the firewall and on corporate websites.
Turns out the rumor was just half correct. Google is not ditching the Mini but it is launching a new service called Google Site Search, which is basically a rebranding and enhancement of the existing Google Custom Search Business Edition.
Google Site Search is intended for organizations that want to add Google search functionality to their websites. It has three main improvements over the former custom search offering: it allows for synonym matching, results tweaking, and indexing of so-called “dark web” content.
If you a run a site that often refers to terms by their acronyms, or commonly uses different phrases to describe things, you can upload a list of synonyms to Google and it will treat them the same when a user searches within your site.
If you’re a news site (like TechCrunch) that tends to place a premium on newer content, you can bias content based on its date, forcing the newest results to the top. Similarly, if you’re an online retailer or provider of some other type of content, you can give preference to results for that content (in the online retailer’s case, any product pages where visitors might go on to buy things).
Google Site Search mainly relies on the standard Google crawler when scouring your site. But you can also hand it special sitemaps so that it recognizes areas of your website that might typically be overlooked by the crawler. These are results like those hidden behind forms or available only through intricate database queries.
Perhaps most compellingly, Google Site Search can be visually customized to match the look and feel of your website. This isn’t a new feature (it existed in the previous business edition) but something perhaps overlooked by most site owners. When you implement site search, you can customize the appearance with a WYSIWYG editor, or even request a raw XML file that can be parsed and displayed anyway you want (you don’t even have to include Google branding on your results pages).
On the other side of the firewall, the Mini is “alive and well” and has even received a recent set of upgrades. Google says that it expects organizations to increasingly use the Mini for only behind-the-firewall type queries (those that retrieve data not available publicly). With the new Google Site Search, there’s no real reason for people to use the Mini to index public websites anymore.
When asked whether Google might eventually provide a hosted solution for indexing private corporate data, company representatives say it’s not ready to announce anything but they “would hope [they] get to that point” eventually. While Google Docs and other online collaboration tools show that enterprises are gradually willing to put their data in the cloud, perhaps the market isn’t ready to hand over the whole basket just yet.
Do you know if Microsoft or Yahoo have anything similar?
Jon
http://dreamclue.com …get the message!
{seesmic_video:{”url_thumbnail”:{”value”:”http://t.seesmic.com/thumbnail/StXHpOOOrg_th1.jpg”}”title”:{”value”:” ”}”videoUri”:{”value”:”http://www.seesmic.com/video/aHrfENphOX”}}}
Woah, scary. But they didn’t release a new search appliance…
Site search is different from web search. One can’t just apply its web search to site searh to meet needs of a site publisher.
Results where not satisfactory when we replaced internal search with google search button for our free technology press release site Google did not do a deep crawl to index all the pages in our site. Recently I read that Google site search will accept “sitemaps”, a xml file with list of links to all pages in your website. Hopefuly this will improve the results returned by Google site search. I wish to replace internal search with Google site search when the number of pages to be indexed exceeds 200,000 pages. Outsourcing the search function to google will hopefuly bring down the load on our database and other resources.
Is this a paid or free service?
Yakov said…
Site search is different from web search
That’s correct. Site search is based on LSI (latent semantic indexing), while web-search is based on PageRank. The difference is that LSI’s input data is a 2D matrix (rows & columns) of document by word-frequency while PageRank is a 2D matrix of inward by outward links frequency to and from a page. So, in short, a matrix of word frequency versus a matrix of link frequency of a document (or web page).
Check out YourAmigo.com
Dumb name, sure, but used by whitepages and other huges sites to make their content more easily indexable and search engine friendly.
They even have a great business model…they only charge based on increased organic traffic.
They also have a couple of patents Google may want to look at
How does the new SiteSearch differ from the FREE Google Custom Search Engine and the not free Google Mini? It seems like these three Google products are competing for the same customer.
Thats very good for webmasters. Are other big search sites doing the same or is it just google?
SearchBlox search is now available on Amazon EC2 (Cloud Compute platform) starting at 35 cents per hour (pay-as-you-go model) and does clustering of search results. http://www.sear...amazon_ec2.html
It is completely web based and gives complete control for indexing of your web sites.
With lots of companies beginning to take a look at the cloud computing platform, this may be where the future is headed.