Roll your own search engine, folks, using Google (Cnet coverage). Dave Pell, I feel your pain….the (partial) release:
The Power of Google Search is Now Customizable
Today, Google is launching the Google Custom Search Engine, a new way to bring tailored search to websites and blogs.
Our clever engineers have found a way to open up the Google search platform to let anyone build their own search engine, without needing a Ph.D. in computer science. In just minutes, individuals, organizations and businesses can use the Google search platform to create their own search engines targeted toward their audience and focused on any content they like, from academic pursuits, to charitable causes, to Hollywood heartthrobs, and more.
When we say we’re letting people build a custom search engine, we mean the whole thing: choosing which pages they want to include in their index, how the content should be prioritized, whether others can contribute to the index, and what the search results page will look like. Custom Search Engines are monetized through the Google AdSense program so they can even generate revenue with it. Universities, non-profits and government organizations can choose not to run ads on their search results if they’d rather not.
You can already see a few Custom Search Engines in action. Intuit’s JumpUp.com site, which provides information and resources to small businesses, is combining a Custom Search Engine with its years of experience in small business to provide the most useful resources on the Web to its users. Or take RealClimate.org, a site that offers expert opinion on the science of climate change. They have created a searchable subset of the Web to provide reliable scientific information to its visitors.
Here’s how a Custom Search Engine works: organizations or individuals simply go to www.google.com/coop/cse and select the websites or pages they’d like to include in their search index. Users can choose to restrict their search results to include only those pages and sites, or they can give those pages and sites higher priority and ranking within the larger Google index when people search their site. Users can then customize the look, feel and functionality of their search engine.
“Search and advertising is at the heart of all we do at Google, and we’re constantly looking for ways to make both even more relevant for our users,” said Shashi Seth, group product manager, Google Custom Search Engine. “Now, people can get the power of Google search, even when they’re not on Google.com.”
Google Custom Search Engine is available at www.google.com/coop/cse . We plan to expand the offering internationally in the coming weeks.
Recall also that Amazon had a version of this with A9.
UPDATE: More from the call today with Marissa Mayer and the product managers. The Custom Search is designed to extend the limited audience of Google Co-op, which they found stagnated as an early-adopter tool. It has improved ease of use, as showcased by the search engine on Justin Timberlake that a Product Manager’s teenage daughter “rolled” in about 10 minutes. While similar to Rollyo’s innovative custom roll, the Google CSE adds the benefit of allowing users to roll an unlimited number of sites together and display the results on their own site, with personalized presentation. Someone on the call described this as the fragmentation of search. The ability to build verticals will allow experts to build specialized engines. But while the engines will be individual, the collaborative element of tagging the domains encourages communities of knowledge to create together. So while each will stand apart from the amazing all-in-one answer box, the Custom Search will also allow a thickening or deepening of intelligent tags in Co-op, which feeds the one box that unites them all.
John – can you get me the link to that Justin Timberlake CSE? I’ve collected some known CSEs here http://wink.com/Google_CSEs–bricktop–61ba6ac9–collections
If you’ve made one or know of one – tag it “google cse” and add it to this list. Would be interesting to see what people are building over the next few days. thx
This was just a logical step for Google to do. It once again shows that Google is developing from a pure “customer brand” into the backend provider of tomorrow’s internet
John,
I just tried the new Google Co-op platform with a list of 162 domains. The search results are poor: My Co-op search engine includes all relevant domains and blogs about web search, enterprise search, desktop search and other search related stuff. Although you can bet that the keyword “google” is included on every URL in this list (on most of them quite often), the search for the keyword “google” led to 65 hits only.
Alexander
The Google custom is neat – particularly the social aspect and the ability ‘to add and label sites in your Custom Search Engine (CSE) or in a Google topic’ using a google Marker bookmarklet (which I assume will be part of the Google toolbar before long…)
Howvewr – is it just me, or could google have been more ambitious?
For example, although subject to similar constraints regarding only being able to search over a limited number of domains, searchfeedr allows you to roll your own search limits from any web feed, delicious account, or web page (in the latter case based on outlinks from a page, or inlinks into it).
Searchfeedr configurations can be saved directly to delicious, and saved searches can, in turn, be pulled into a page.
Searchfeedr is an example of a search intermediary in other ways too. For example, domain or page limited domain searches can be run over different search engines – at the moment the options include Yahoo, Google, Windows Live and altavista. If no searchfeeder limit source details are provided, you can just search the search engine normally.
Currently, actual search engine results pages are used to display search results, although I do have some versions with embedded search results brought in via the search engine APIs. However, while testing the service embedding the actual search engine results pages clearly demonstrates that searchfeedr can obtain page and domain limited results from a variety of search engines.
I wish there were something between “anyone” and “invited individuals” for collaboration. The former opens the door to web spam, which will soon come in. The latter is too simple a model for building a large community.
Good that we (Microsoft) already have such great followers! Keep it up Marissa:)
I just made three different search engines as an experiment. What I found out was that Google COOP can’t search subdomains of sites, such as fas.harvard.edu, unless speficially mentioned.
I am definitely disappointed with the service as is.
But my Web 2.0 Blog Search Engine had a little more success as most of the sites did not rely on subdomains.
Two features seem to make this vastly superior to LIVE and Yahoo (on which Rollyo is based): allowing hundreds of URLs and allowing you to “boost”, rather than “only include” your chosen URLs.
Still, I find it interesting how HUGE the “NEW Google thing” buzz is for a search product that, although better that the predecessors is not really new at all.
Google sucks….I think folks are getting so sick of these guys. They are like the evil empire. Remeber folks, no matter how much money Google spends on PR trying to convice you and journalists that they are different, new, innovative, thoughtful and smart….they have carefully planned and executed a strategy to dominate the web to make BILLIONS. Congrats google….Sorry rollyo
Joseph Hunkins, you should go and check out Live Search Macros. We have some really powerful constructs available. We have a construct called, “prefer”, and there are other constructs also to sort of emphasize a set of websites, “link**” constructs. Now go to Google’s custom search engine site. Compare the available constructs and their working yourself.
If you are at it, then also, check out our advanced search feature.
It’s definitely a great news from Google. Will it be a big blow to search engine companies like Trulia.com or Guruji.com that tacke specific verticals ? They have invested lot of money and research in building such engines but now anyone can make similar engines free using the Google co-op custom search engine.
“Google CSE adds the benefit of allowing users to roll an unlimited number of sites together…”
John, are you sure about the unlimited bit? Annotations per co-op account on Google seem to be limited to 5000 annotations across all CSE’s one has – but I haven’t yet seen any evidence of there being NO CAP on the maximum number of annotations *across* all contributions to a CSE.
If that is indeed the case, then well, more power to Google CSE users.
Regards,
-Nas
I’m finding that the results are somewhat hinky. It’s documented that subdomains have to be specified, but it seems also to be failing to find pages in top-level domains, pages that are found when doing a search on the main Google page.
Some nice features, but I wish it, you know, worked more like Google.
Almost one year on from the last comment by Shannon little has changed. Google’s CSE is still not finding pages that are found by the main google search. Sometimes CSE only finds 10% of the pages that main Google search finds on the same (obviously specified, with correct subdomain) site