Google Search Seeks Speed

Amit Singhal discusses the future of search at the SABEW conference./Chris Prentice

By Matt Robinson

Google is synonymous with search, but could be onto to something greater: inference. Amit Singhal, Google Fellow and tweaker of Google’s search algorithm since 2001, says search is beyond key words and at the door step of understanding language. Singhal, speaking at the annual Society of American Business Editors and Writers conference in Phoenix held in late March, used some examples. If you google “GM cars,” you get General Motors, but if you google “GM food,” you get genetically-modified. “Search is about matching meaning to what users want,” he says.

Another example, if you google “panasonic lock TV,” you get hits about “parental control.” Singhal says marketers probably didn’t like the word “lock” and went with the nicer “control.” But Google infers the difference. Other points made at the conference include:

• Information wants to be free he says, but that doesn’t mean websites shouldn’t impose a pay wall, citing that good content costs money. But differentiating that good content from “McContent” - quick, cheap generated content - is difficult since the language between the two is often similar.

• Web content is suppose to be free. And citizens with information are better ones. Singhal sees China inhibiting its best interest by restricting search.

• The future of Web search is still based on the what made google popular: speed. “We are crazy about speed,” Singhal says. He noted that if a user’s is slowed down by 200 milliseconds, the site will end up losing traffic. Relevant algorithms are now more important than ever, he says. “As the amount of information has exploded, the importance of relevance has gone through the roof,” he says.

• The kind of information that people seek on phones is much different than what people search for on their computer. Singhal sees a future in location-based advertising.

(Images from home page are courtesy of the Tuscon Citizen.)

UK Launches Semantic Data Site: Will the Rest of the World Follow?

By Linda Broughton

Sir Tim Berners-Lee, credited with conceptualizing if not inventing the World Wide Web, is not finished yet.

Lee and Professor Nigel Shadbolt, both appointed as Information Advisors to the UK Government, are coordinating Data.gov.uk. Data.gov.uk plans to provide the general public with a single access point to all the United Kingdom’s national, regional and local statistics, surveys and studies - all the information in the UK collected under the umbrella of the national government.

In a passionate speech at TED, Berners-Lee explains that no data is an island - it’s the relationships between different data that make the data valuable, insightful and useful. And these relationships are not often visible to the human eye - if anything, our own preconceptions and individual personalities often interpret rather than understand the relationships between different data, closing us off to real social trends and concerns.

Berners-Lee wants to use the Web to bypass the human attempts to dress up data for personal or political purposes. Instead, he wants the Data.gov.uk platform to expose the real significance buried within the data-to-data relationships through encouraging the digital mapping of raw, unadulterated social, political and economic data. Berners-Lee calls this ‘linked data,’ a precursor to the semantic Web that he believes will one day explain the meaning behind how we collect, calculate, evaluate and use the data that we post online.

Not all of these data relationships will be obviously significant - imagine an application that maps the relationship between data discussing the annual amount of Cadbury chocolate sold in Yorkshire and the corresponding annual number of failed marriages within the same geographical location and time period. However, the project is expected to help identify unexpected and insightful data relationships, insights that would normally take several decades and hundreds or thousands of brilliant socialist scientists, statisticians, psychologists, focus groups and public policy experts to simply suspect.

The automatic data relationship mapping will allow the UK government and the UK public to discover connections, trends and causal relationships that will inform public policy for decades to come. If implemented properly, Data.gov.uk could come very close to comprehensively mapping and explaining the past and real-time behavior of the public - and thus allow platform users to accurately plan the future of a society.

Of course, the key to the concept is the raw data. Currently, there are already third-party applications that map roads and potholes in the UK, provide statistics and information about the location of doctors throughout the UK, and give up-to-date information about local schools. This is useful data to aggregate but it is not yet generating anything that a few quick, targeted searches online can’t. The next push will encourage interested developers to create applications linking the different data to generate data relationship maps that give researchers at think tanks and academic institutions something to ponder and investigate.

Supporters of the open data movement are urging private businesses to follow the UK government’s example and release their raw data to the public. If the private sector keeps its data too close and too secret, the sector risks losing potential profits that would arise from information generated by an external comparison and review of their aggregated data. Moreover, the private sector makes up an important part of modern society. Accurately identifying, explaining and impacting public trends requires more than the government’s analysis of the population’s behavior, it requires an accurate understanding of the public’s practices in commerce and industry.

Data.gov.uk may soon be the way for the UK government (and anyone else interested) to keep several fingers not only on the pulse of modern UK society, but on its stomach, windpipe, eyes, mouth, ears, etc. The platform and its third-party applications may soon provide an in-depth and automatic monitor of modern British, Scottish and Northern Irish daily public life. Do the industries want to jump in now, developing their own applications and supplying their own data to complete the public picture, or will there always be a yawning gap in the data buried in the private sector’s own digital databanks?

http://www.opte.org/maps/tests/

A Mind of Its Own: Search Engine Technology Ever Pervasive

By James Zipadelli

Americans performed more than 15 billion searches in January, which is up 3 percent from the month before, the audience measurement service comScore says. The latest search engine rankings show that Google is still king when it comes to search engines. “Google Sites accounted for 9.9 billion searches, followed by Yahoo! Sites (2.6 billion), Microsoft Sites (1.7 billion), Ask Network (574 million) and AOL LLC (375 million),” the release says.

Although Google spokesperson Nate Tyler declined to comment on Google’s numbers, he did say that Google Suggest Technology is an effective way to help users search for what they are looking for.

“As you type into the search box on Google Web Search, Google Suggest offers searches similar to the one you’re typing. Start to type [ new york ] — even just [ new y ] — and you’ll be able to pick searches for New York City, New York Times, and New York University (to name just a few). Type some more, and you may see a link straight to the site Google thinks you’re looking for — all from the search box,” Google’s Help Forum says. (Ask.com and Microsoft were not available for comment at press time.)

Kevin McFall, co-founder of the vertical search engine RushmoreDrive.com, says the level of difficulty “is pretty high” for anyone trying to gain a share in the search engine market because established search engines spend large amounts of money on marketing and advertising. However, he says there are ways a new search engine can differentiate themselves from their competition. RushmoreDrive.com was a sister site of Ask.com and was shut down in June 2009 due to the recession.

“One must position the value of one’s search in such a way as to change existing behaviors and habits of those who already use Google, Yahoo, AOL or Bing by offering them a reason to change and then delivering a rich enough experience to warrant their frequent return,” McFall says. “One must also realize that instead of taking on the major search players head on, one must find a way to backdoor them to get a slice of the market share instead of trying to compete directly.”

According to McFall, he was able to do this with Rushmore Drive.com by marketing his website as a discovery engine and a search engine. “We achieved the ability to deliver a richer and more relevant set of results through our unique index and page ranking algorithm, along with a distinguished universal results page, which delivered text, image, video and blog results all in one page,” McFall says.

He also suggested search engines that have a social component would be more successful long-term.

There are also specialized websites that find search engine technology useful. For example, Healthline Networks uses search engine technology to help customers with health and drug information.

Healthline Networks CEO West Shell says, “We’ve found out that consumer search can be complicated when it comes to health. Consumers and doctors speak different languages, and often consumers don’t know what to look for when they start.”

Shell says the technology Healthline Networks uses is based on “semantic taxonomy,” or classification, of health information. He also says the technology is always being updated to ensure customers have the latest information available and that they are partners with health carriers like Aetna.

Rich Kahn, CEO of the search engine eZanga.com, says his search engine is being redesigned and should be finished by late 2010.

The redesign allows eZanga.com to “significantly increase the number of sources we pull information from, improve our relevancy algorithm so that our results will be more accurate to the queries performed by our users [and] designing new technologies, that are not used by any other search engine at present, that will improve how we display our results to users in a way that will be more useful to our users,” Kahn says.

Is Real-Time Search Working?

By John Greaves

Real-time search lets the user seize timely information from a quick-flowing stream of data. According to analysts Matt Booher and Ilona Vijnik, “It’s a step toward giving users the ability to access highly relevant and fresh information, delivered in real time from across the Web.” From the proliferation of real-time search engines like Wowd and Collecta to Google’s recent entrance into real-time search, it seems companies are leaping at the chance to tell us everything immediately.

Booher and Vijnik both of Bridge Worldwide, a Cincinnati-based digital marketing agency say the main question to consider when asking if real-time search is working is to consider how the information provided by real-time search furthers a specific objective.

This includes users who tracked the development of the Kanye West-Taylor Swift story on Collecta and those who shared YouTube videos through OneRiot after learning of Michael Jackson’s death, as well as those who followed weightier issues.

“What’s happening in Yemen right now you can’t find through traditional search,” says Gerry Campbell, CEO of Collecta.

So is real-time search succeeding? According to Tobias Peggs, GM of OneRiot, “Our definition of success for real-time search is ‘Is it adding user value?’ and from a business perspective can that user value be monetized?”

One monetization model involves having clients pay for real-time services. That’s the model used by Jobvite, a real-time search recruitment tool, which boasts it can give job recruiters a real-time picture of prospective candidates based on their total Web presence. Jamie Glenn, chief products officer at Jobvite, says real-time search is crucial to the company’s success. “Recruiters can use our tech to go across social Web to find people who fit a job to see if they’re interested. Once they find that person on open Web or social network they can see that person’s profiles on Facebook, Twitter and LinkedIn.”

Adam Hyder, CTO of Jobvite, says using real-time search enables recruiters to begin to build relationships with candidates. “We had a recruiter who saw a candidate’s wish list on Amazon and he sent a book from the wish list as a gift to that candidate,” Hyder says.

However, not just vertical search engines want to monetize real-time search. According to OneRiot’s Peggs, monetization makes relevancy crucial to making real-time searches more than a novelty item. “If you’re delivering a fire hose of information with no relevance attached to it, it’s kind of hard to figure out what advertising should go with that,” Peggs says.

Analysts Booher and Vijnik say current hot searches include whale wars, Richard Blumenthal and Nexus One for Verizon while hot topics include Microsoft Office 2010 and Nexus One Google phone. This is important data for marketers who use it to tailor ad campaigns to popular searches.

Sites like OneRiot are actively trying to monetize those searches and willingly collaborate with marketers to do so. “If you’re a developer you can grab our API and use our real-time search results from it, we provide tech support, we in addition offer real-time ads, if you’re a developer who’s proved your users like real-time ads we can do a 50/50 revenue share,” Peggs says.

Monetization requires consumer loyalty. Mark Drummond, CEO of Wowd, says providing relevant information brings users back to your site. “People want to be engaged with material events as they occur. A lot of search engines say we’ll vomit at you all the tweets with these words in it. If I want tweets, I’ll go to Twittersearch. ”

Unfiltered searches not only detract from relevance but also attract spammers according to David Evans, CTO of deeplocal.com, which produces the newspaper search engine NewspaperNinja.com. “The problem here is mostly dataset. If we’re seeing a hard time detecting spam in email, we’ll see a really hard problem for Bayesian detection algorithms at about 20 words to sample from. Especially when we’re abbreviating everything and jargon is changing at an alarming rate,” Evans says.

Add to that tweets from disgruntled customers according to Dave Conklin, president of Internet Marketing for ProspectMX. “As human beings we don’t tweet when something is working. The second something goes bad, you tweet about it and you get a skewed point of reference in real time,” Conklin says.

Some sites fight spam by ranking links based upon user feedback. This is the approach used by Jobvite and Wowd. OneRiot’s approach indexes results based upon investigating links to see what’s behind them. Various websites have documented Google’s real-time vulnerability. Google acknowledges the danger from spam but a Google spokesperson says, “Google search aims to show users the most relevant results for a given query. We apply the same high standards to real-time that we do to the rest of the web. There are always unscrupulous people who will try to game our ranking systems, but as always we are uniquely equipped to suppress spam content.”

Spam is also an issue for Internet marketers who post multiple times to keep up with random tweets. They have to be careful to avoid looking like spam. Bob Bentz, who monitors placement for Advanced Telecom’s mobile marketing product www.84444.com, says unstable rankings are behind multiple postings in Google, Bing and Yahoo on many of their company’s most important keywords. “Google rankings are changing sometimes several times per day! Yahoo is not changing daily and Bing rarely changes, although more often than Yahoo does,” Bentz says. “What it says to me is that we need to continue to send new content out there several times per day to keep our rankings active on the first page.” Bentz hopes filtering systems such as Wowd’s panel rank can differentiate between robotic postings and original content.

“I think sometimes we’re concentrating on SEO and we don’t see the forest for the trees. If your stuff is seen in blog posts and articles and someone sees it and wants to come buy your product that’s the goal, not ranking,” Bentz says.

Profitability is what the future of real-time search is about, according to OneRiot’s Peggs, “2009 was about proving that real-time search adds value. 2010 is about monetization.”

How a Search Engine May Choose Search Snippets

By: seobythesea.com

When you search at Google or Yahoo or Bing, you’ll see a set of search results that include a page title, a summary or snippet of the page, and a URL indicating the address of the page.

[Read More and Discuss]

Video SEO vs. Traditional SEO, Misconceptions & Opinion

By: reelseo.com

A few weeks back, I was asked by Jeff Boudier, VP Business Development for Stupeflix, if I would be willing to answer a few questions about Video SEO for their blog.  You can view the whole interview here, but I thought that I would like to provide a few of my thoughts to ReelSEO readers.  Id love to know your thoughts.
[Read More and Discuss]

Next Page »