When you’re the distant second player in web search, you’ve got nothing to lose by making bold moves. So it makes sense that Yahoo has adopted an open strategy with the following idea in mind: woo developers to build on top of your technology, and then display your advertisements to more eyeballs throughout the long tail of the web.
Yahoo takes a large step in that direction today by announcing a radical and historical initiative called BOSS, which stands for “Build Your Own Search Service” and basically turns web search into a web service by inviting developers to leverage Yahoo’s core search technology and build their own web search implementations. Hakia and Me.dium are among the first to switch over and use Yahoo to power their web results (in Hakia’s case, as a supplement to its own search technology).
BOSS allows developers to submit queries (and their associated parameters) via an API to retrieve up to 50 web, image, news, or spelling results in XML or JSON format at a time. Per Yahoo’s policy, developers will be required to display its ads next to, or within, their results (although this requirement won’t be imposed until later, Yahoo plans to offer CPM fees as an alternative, and academics will be exempt from any such attempts at monetization completely). Yahoo won’t require anyone to mention that they are using its technology, as it doesn’t intend to drive traffic back to Yahoo proper - just spread its technological influence throughout the web. Nor will it impose any caps on daily queries, or impose standards on the design of results pages.
The self-service BOSS API can be used directly or via a mashup framework that simplifies the process of aggregating data from across the web. When using the API, developers can choose to filter porn from web results, specify the dimensions of image results, and retrieve only news results from within a certain number of days. The order of Yahoo’s results, however, can only be changed after they are retrieved (you cannot pass Yahoo parameters that tweak its standard relevancy model and return reordered results).
If you want to incorporate ancillary data with the results from Yahoo’s main index, the Mashup Framework simplifies the process by which developers conduct joins on web data (in formats like XML, JSON, RSS, or RDF). For example, it can be used to generate results with information from both Summize and Yahoo Search, Digg and Yahoo Search, or pretty much any API and Yahoo Search. The framework will work only through Python constructs to start, but more languages should become supported in the future.
All of this forms just the first half of Yahoo’s intentions with BOSS. In a few months, Yahoo will also release APIs for what it’s calling “BOSS Custom”. This version of BOSS will allow developers to actually push data to Yahoo’s servers for indexing, and then perform highly customizable search queries against them.
So in a way, BOSS starts by opening up web search to 3rd parties, but it will go far beyond that by providing cloud-based search for all imaginable types of data. Yahoo has already enlisted a handful of partners for BOSS Custom, including us here at TechCrunch. We’re working on a search implementation that will enable readers to conduct searches across the entire network and retrieve results that have been weighted using a custom relevance model. Readers will also be able to drill down by author, comments, date, and other criteria.
BOSS is the second concrete product to come out of Yahoo’s Open Strategy. The first was SearchMonkey back in April.
Congrats to the Yahoo BOSS team! From all of us at Me.dium.com, it was a pleasure working with you. More (much more) to come…
Outside of the few techies in the Valley, i just don’t see the application of this. This Boss deal is like putting lipstick on a pig. The fundamental problem with Yahoo search business is their “useless results” so now they are exponentially making their useless results more useless.
I don’t see the widespread application of this. When i talk to Real people (not techies), everyone uses google.com So all these services springing up with their flavor of search seems like absolute waste of time.
What happened to real innovation not the regurgitation of the same bull.
Just my thoughts.
Wow - this is what I would like to see come from Yahoo. Excited for more to come.
Correction from hakia:
We are not replacing our own semantic QDEX system (hakia’s process of indexing the Web) with the Yahoo Search BOSS . hakia’s participation in Yahoo Search BOSS means that we now have access to one of the largest Web directories on the Internet, which greatly accelerates our QDEXing and semantic analysis of the Web’s content.
Sorry for the inaccuracy, Farrah - I’ve updated the post to include your correction
Boss is an unfortunate name.
Thanks Mark!
Getting my dev tools out.
I like the logo.
Sorry, I don’t want to be totally negative. I think this is a terrific initiative for Yahoo and partners.
BOSS was a straight forward fit to compliment Daylife’s news search pages with Yahoo’s web search results. Much more to come - http://www.daylife.com/page/yahoosearchondaylife
- Vineet @ daylife
Google has provided this for a LONG time now, including the “push” custom indexing
http://googlecustomsearch.blog.....-apis.html
BOSS sounds a lot better than BYOSS :-).
this move makes accelerates search innovation around end-user experience and hurts google where they’ve been lethargic. it also undermines search’s standing as a destination-driven business.
i posted about searchme.com just yesterday, and relevancy and index size are the only impeding factors for them. now there is no stopping for them and other niche players to get good traction.
well done, yst.
yawn. and how many of you use searchmonkey? does anyone?
found an article on google reader just now written by someone who worked on this project:
http://zooie.wordpress.com/200.....ider-view/
A question : Will Yahoo provide a Custom Search Engine type of feature which Google is currently providing?
I don’t want to sound stupid with the above question, but the idea is to ask if an average non-techie be able to use this facility
Hey all you people smarter than me:
1)How is this different from Google’s CSE program - http://www.google.com/coop/cse/
2)My experience with Yahoo’s APIs have been shoddy at best. I read press releases touting how open and accessible their services are, but when I try to access and utilize their tools I often run into problems. Either they’re not sufficiently documented or I get something about them besing reserved for key “partners”…which is rarely me. Would love to hear if the massess can truly take advantage of this thing. Could be neat.
ditto
Tech: I’m sure someone will make Wordpress, Drupal, IFrame, etc. plugins for turnkey users.
This is a great move by Yahoo, but no quote from Arrington? Surely this is another sign of their ill health and lack of leadership and vision.
Can’t wait to see what Yahoo would open up next with it’s Open Social initiative. Opening up it’s search index is one thing, whereas opening up 600 million users worth of data would be of epic proportions. Go Yohoo!
can someone outline the main differences between GOOGLE API and the this new YAHOO API?
Will they just offer more search results?
because at the moment Google only offers up to 4 pages worth of results with their API.
If Yahoo can offer more than 4 pages than that is awesome.
Tim,
There is no daily quota for the searches. That is the main difference. In this sense, you can provide service on your site and don’t need to worry about using up the quota.
but ther eis no qouta for the google api either.
Only resrticted to 4 pages worth 8 search results per page.
What is the diference between BOSS and the old Google API?
sorry about the spelling mistakes i mean
4 pages and every page has 8 search results
so that is 8X4 = 32 results per keyword.
Yahoo will have more?
@Tim Yahoo will let you fetch 50 at a time and keep going all the way down.
Thanks John,
I wonder if Google will refute back by offering more than 50 results lol.
Let the war begin
The Google api is extremely restrictive. At least for the SOAP search api, it’s for personal, non-commercial use only. They limit the number of queries per key, and each organization is allowed just one key.
I think their adwords api also has some horrible restrictions on how it can be used, too.
I’m not quite understand with the following sentence:
“The order of Yahoo’s results, however, must be changed after they are retrieved (you can’t pass Yahoo parameters that tweak its standard relevancy model).”
Does it mean the order of the results from Boss (pages, images…) will be different from the results from yahoo’s search page???? Therefore the sequence of the results are somehow classified???
Wayne, I believe what the convoluted sentence means is that if the developer wants to change the order of search results, they may do so (as contrasted with SearchMonkey add-ons, which may not change the order of search results), but if they wish to do so, they must do so after retrieving the results, not on-the-fly as part of the querying process.
We just developed an interesting SearchMonkey app (see http://shoppingnotes.wordpress.....ng-monkeys).
Now BOSS opens up even more of Yahoo!’s search platform. Looks like Yahoo! is really serious about this open approach. Bravo!
@Zizzle you obviously lack the talent to read and comprehend English.
This is not radical at all. Just more mashups, with no true support for vertical search engines, which require new parser/classifiers for indexer and searcher and potentially new index format for ranking support to do anything novel.
It’s just another API for their existing indices/services. No fundamental improvement can be made from this approach.
Do you think a thousand mediocre search engines with different skins can fight a better Google?
Just open up a sizable portion of their cluster a la EC2 with a reasonable fee but with each virtual instance preconfigured to access to their entire crawl data for people to pick and choose to build their stuff. Now that would be radical and disruptive. Potentially profitable as well.
A Yahoo search API for their various services existed for a long time… BOSS seems like a new name for it, with a couple of extended features (like allowing to reorder results, though people could do that before, too).
Excellent… and high hope of it.. keep it coming!
Its a step in the right direction.
Although Yahoo is “the distant second player” They still have much to lose. All this does is create some much needed positive buzz for Yahoo and put a little bit of pressure on Google.
Shopping Notes.com: I saw your application and it looks interesting as I can see the prices of articles immediately. However, at many places in the search result, it shows ‘Enhancement Failed’..Why is it so?
Yahoo BOSS team is great. We had an easy time implementing the API in cluuz.com. This changes the landscape for alt search engines.
Great job and thank you.
Opening Yahoo to yahoo results is just that..yahoo results. People use google for google results. My 8th grade football coach once told me…as I was first and lazy across the finish line..”Armstrong…just because you are first, doesn’t mean you can be lazy”..yahoo was first and got lazy.
Multi-site..multi vertical aggregation with a REST/JSON/AJAX api is already done…www.bouncewebservices.com
Is searchmonkey a thing of past now?
So Yahoo wants me to build a search engine piggybacking on theirs and they can use it to display ads and make money not for me, but for them? Am I understanding this correctly? Why would anything seriously get built?
I’m going to use this for the home page of my social site in a few weeks.
This is really cool. This opens lot’s of possibilities to rerank results based on del.icio.us , other social bookmarks etc.
The main difference with Google CSE is that Google requires you to use their logo and don’t allow you to mix your results from other search engine providers and you cannot monitize other then with Adsense. XML output is not available unless your have a paid business CSE ( but these only search 5000 sites). This innitiative looks more open/flexible.
Huge potential on BOSS. All type of “on the fly” content generation can be implemented.
Will push Google to open up their API
@scoofers saw your search engine. think you’ll swap out?
> filter porn from web results
There’s more money in the reverse, filtering the non-porn to return “all porn” results.
@grm
We will certainly test BOSS, but Google is also not sitting still. Their CSE team was very busy (its my impression) with rebranding the CSE to Google Site Search recent months. I hope they launch some new features soon.
What came today is the first half of exciting. The second (bigger) half of exciting is the BOSS Custom framework. If that really allows companies to add data to Yahoo’s index, it could generate a win for Yahoo by growing the size of the whole search market.
For example, say some company has an engine which tags telephone numbers in web pages, and provides a link for each allowing the user to call the numbers using Skype (or similar). Previously, such a service would be limited to that company’s website or portal, and perhaps a few affiliated sites. Now, the service can be linked in to Yahoo Search directly and present those links with some or all search results. Yahoo, the company, and even Skype, get a bigger slice of the (bigger) market.
And when you’re a distant third in search (Microsoft) you use Corporate raiders to fake a hostile takeover, and other borderline illegal actions:
http://www.techuser.net/microsoft-yahoo.html
Seriously, from reading this commentary on Yahoo you’d think that markets can’t have more than a single player. TechCrunch’s understanding of capitalism is impressive, or is this another pay-per-post shill for Microsoft?
BOSS Custom sounds a lot like Google Site Search - costs money, you can feed it data, you can filter results, results come in as XML, etc. Perhaps I’m missing something… Note that Site Search is a bit like Google Search Appliance but no hardware and a lot less flexible.
Medium?
Hakia?
Techcrunch?
Utter failure.
Too niche to ever matter to Google. What does this do to lijit?
Google can change their API at will. Change the game, don’t play off of it. I’ll even spoon feed it to you: VRM
sounds like this release is a search api that’s totally unrestricted
change the presentation, re-rank, no attribution
google can do this with theirs quite easily (TOS change), but if this puts pressure for them to do this then great
also, google cse lets you search over restricted results and do refinements
this api doesn’t do that - it’s a straight up REST search api
reading the boss custom page, it sounds very open
blending machine learning models, exposing ranking features
i think they are really trying to be super open here
what i want to see is open search code
maybe that’s the next step
regardless, a good simple first step.