Bigmouthmedia search engine news: algorithms

Google announces new Image Search details

Wed, 30 Apr 2008 12:00:00 GMT

Internet giant Google has announced that its new image search algorithm will make finding images simpler.

During a web conference in Beijing, the search engine provider unveiled a new tool which is designed as a means of ranking images in order of importance.

According to the Times, image search has presented a problem for technology firms because search engines are not yet capable of deciphering objects within pictures that would be instantly recognisable to humans.

However, Google believes that it has devised a more effective approach to the problem.

The tool used in Google Image Search works by recognising a particular visual theme or pattern in a series of images.

Its relative importance is then analysed on the basis of how similar it is to other images which also contain the theme.

The example cited by the Times involves the golden arches associated with McDonalds.

According to the publication, Google Image Search would draw up a set of links between images containing the familiar golden arches and rate those which showed the "typical yellow M" the most highly.

Commenting on the new offering, Google researcher Shumeet Balujaz stated: "By treating images as web documents and their similarities as probabilistic visual hyperlinks, we estimate the likelihood of images [being] visited by a user traversing through these visual hyperlinks."

A team of 150 Google employees has been working on the creation of the scoring system for the relevance of images.

Digg This! | Save to del.icio.us | Reddit | Submit to Netscape | Google Bookmark This | Furl This | Add to Yahoo MyWeb 2 | Contact bigmouthmedia

Flaw in White House web strategy exposed

Thu, 31 Jan 2008 12:00:00 GMT

The White House is in danger, and this time the enemy comes from within.

A bigmouthmedia technical report has identified a potentially crippling technical weakness at the heart of the US Presidential office's internet strategy. If not addressed, the critical error threatens to have an insidious effect on George W. Bush's campaign to win the hearts and minds of the online public.

In a report published today, our in-house analysis team reveals that the official White House website's robots.txt file - which acts as an index guide to the site - is becoming so large that the welter of information published by the Oval Office is in danger of disappearing from the world's search engine results. If unchecked, the crucial file could soon reach the point of no return, exceeding Google's maximum cut-off size of 100kb and becoming so large that it may be inappropriately terminated by the search engine's technology.

Crucially, the bigmouthmedia report warns that if the maximum file size for whitehouse.gov's news directory is exceeded, an entire range of content published within it could be blocked from the majority of search engine spiders.

Vital tech issues, such as net neutrality and technology education, have been playing an unprecedented role in the run-up to this year's US Presidential elections, with tech sites and industry blogs profiling each candidate's credentials on IT issues. In fact, in 2007, Hillary Clinton, Barack Obama and John McCain all gave talks at the Googleplex, Google's company headquarters in Mountain View, California.

What's more, influential tech blog Techcrunch gave Obama and McCain their endorsement for "Tech President" live on Fox News on January 29. But which candidate will be able to save the White House website?

Andrew Girdwood, Head of Search at bigmouthmedia, commented:

"The White House bans search engines from some parts of their website but allows them to read and index content from other parts. Problems will arise when the 'banned list' gets too large. The search engines will not be able to cope and Google's been known to truncate instructions midway through. When that happens, a whole section of whitehouse.gov could vanish from search rather than just the pages the White House intended to block."

Read the full report: "Crucial White House website flaw uncovered"

Digg This! | Save to del.icio.us | Reddit | Submit to Netscape | Google Bookmark This | Furl This | Add to Yahoo MyWeb 2 | Contact bigmouthmedia

New Google Sitelinks Update?

Wed, 17 Oct 2007 12:00:00 GMT

Now officially the most popular search engine on the planet, Google appear to be testing an upgrade to their selective Google sitelinks algorithm.

The search engine has expanded the sitelinks selection that floats under the main site URL to eight quick links instead of the usual four. Designed to help users navigate a site more effectively and originally released in September 2006, the sitelinks process is an entirely automated algorithm.

Google analyses the link structure of the site in question and finds 'popular' shortcuts that save users time by finding information and exposing useful links within a site. Typically this information might be contact details, about us pages and product listings. But due to the automated nature of the system, Google may choose to select other content in the site as a prime candidate for the sitelinks treatment.

Whilst the new improved sitelinks algorithm appears to be simply in testing phase, the results seem to appear both in organic search results and Google Mobile. With the recent upgrade frenzy that has caught all the major search engines, particularly Yahoo! and MSN, its likely this new improved algorithm will be rolled out with appropriate fanfare sooner rather than later.

Digg This! | Save to del.icio.us | Reddit | Submit to Netscape | Google Bookmark This | Furl This | Add to Yahoo MyWeb 2 | Contact bigmouthmedia

World's leading search engines announce Sitemaps Autodiscovery

Thu, 12 Apr 2007 12:00:00 GMT

The world's search engines, including Google, Yahoo!, Microsoft Live Search and Ask, have announced they will be supporting 'autodiscovery' of sitemaps. This means that webmasters will be able to specify the location of their sitemaps in the robots.txt file.

Thus with the new format, sites will no longer have to submit their sitemap to search engines separately. The following line should be added to the robots.txt file: Sitemap: http://www.example.com / sitemap.xml

This new autodiscovery format will benefit both webmasters and search engines. They'll save time for webmasters, as they will no longer have to submit their content to search engines.

Search engines will automatically acquire information on pages to index, content and meta data, updated pages and the most important pages, all from the robots.txt file. The sitemap protocol at sitemap.org is now available in 18 different languages so that sites globally will know about the new sitemap format.

Ask.com, Google, Microsoft Live Search and Yahoo! are all supporting the new autodiscovery sitemaps format. Ask.com is also now supporting submission of sitemaps through http://submissions.ask.com / ping?sitemap=SitemapUrl.

Like manual sitemap submission, autodiscovery will not necessarily index all site pages. The pages are still required to meet search engine criteria in order to be indexed. In the early stages, it may be worthwhile to manually submit sitemaps to ensure they will be processed. This will enable webmasters to have access to specialized monitoring and reporting tools, or information on how they crawl your site.

Digg This! | Save to del.icio.us | Reddit | Submit to Netscape | Google Bookmark This | Furl This | Add to Yahoo MyWeb 2 | Contact bigmouthmedia

Google release revamped mobile search

Wed, 28 Mar 2007 12:00:00 GMT

Following Yahoo!'s recent release of their new mobile search service, oneSearch, Google have seen fit to revamp their own mobile search service. The new system has been in the testing stage for a few weeks; but bugs have been ironed out and it has been unleashed into the mobile world.

When your mobile phone is pointed towards "http://mobile.google.com" what you now see is a mobile-enhanced version of the Google homepage, with links to search for a term or to build a personalised homepage for yourself.

Google's aim, as laid out on their official blog, is to provide the most relevant services for its users while they are on the move. Google Account holders can log in to their homepages and view personalised gadgets and feeds, including weather, stocks or news, all laid out to fit onto a smaller screen. Users can add, re-order or edit all of the gadgets just as they would on their traditional Google homepage.

The mobile search algorithm has also been tweaked for Google Mobile. The algorithm now recognises if you are looking for web sites, images, local information and other similar information. If you were to search for "pictures of dogs", for example, the algorithm would automatically realise that you are probably carrying out an image search for "dogs" and so would promote image results to the top of the results page.

Similarly, if you searched for "McDonalds", the algorithm would show you the addresses of McDonalds restaurants in your default locality at the top of the results, along with a link to a map highlighting McDonalds restaurant locations.

Other small alterations made to the new Google Mobile Search service are the merged mobile and web results. Pages specifically created for mobile phone viewing will have a small phone symbol next to their link. Additionally, instead of cycling through a whole page to find what has been searched for, Google will automatically show the relevant portion of the webpage.

The calculation function, as well as some fact functions, are noticeably absent from this reincarnation of Google Mobile, which means searching for gems such as "2386 furlongs in lightyears" can't be done on the move. But it's highly possible that Google may well implement these features in the future, just as they have with movie search information.

Digg This! | Save to del.icio.us | Reddit | Submit to Netscape | Google Bookmark This | Furl This | Add to Yahoo MyWeb 2 | Contact bigmouthmedia

Google launches CPA model for advertisers in the US

Tue, 27 Mar 2007 12:00:00 GMT

There has been many a long discussion in the internet marketing world as to the direction in which Google will evolve its paid advertising format. On Tuesday 20th March 2007, Google gave us a further clue about their plans as they launched a CPA model for advertisers in the US. The model has undergone testing since the summer of 2006, but now a beta format has been released. This switch to beta is worth noting as, with many of Google's new products, it indicates that the CPA model is heading for a full release in the near future.

Google's CPA model allows publishers to display adverts and be paid on a completed action basis. Previously, Google has limited this form of advertising to a 'per click' (Adsense), or occasionally a 'per impression', basis. It allows the advertiser to set a fixed amount that they are willing to pay per action - or your cost per acquisition (CPA). Importantly, advertisers do not need to be signed up to the content network to operate CPA ads.

Some immediate limitations have been identified by affiliates and publishers, notably the stipulation that publishers can not use PPC to promote pages where Google CPA ads are displayed - but the extent to which this is enforced by Google remains to be seen.

The move has been labelled by some as a direct threat to affiliate networks. However, Google will not provide certain key attributes that are provided by affiliate networks. These are:

Exact commission amounts - similar to Adsense, publishers will be paid within a fixed range.
Tailored adverts - CPA ads will keep the Adsense format.
Choice of advertiser - Google will determine who is shown on publisher sites, unlike an affiliate network where affiliates can promote specific merchants.
Support - a USP of most affiliate networks is that they provide advice to affiliates and publishers to help achieve conversions for the merchant.

If Google gets close to this model of support and relationship building, they lose completely the self-service, hands-off approach that has allowed them to access the mass-market.

However, this is also an argument that may see publishers preferring to remain pay per click. Why would a site with a decent source of income via Adsense, risk the switch to the CPA ads only for the merchant sites to convert poorly?

For advertisers, Google's new CPA model provides another measurable tool to test. There will be almost unlimited potential for advertisers to format offers mixed with variable CPA base rates. It should not be confused with a pure affiliate program, however. These are still just 'ads'. They are not banners, coupons, vouchers, cashback incentives, email templates or any other resource that allows affiliates to integrate a merchant into their site.

If Google can now be called an 'Affiliate Network', then it should have been given this name many moons ago when it launched Adsense, a model which many view as a pay per click affiliate program.

Google's CPA model may provide an immediate threat for smaller affiliate networks in the US, but for larger organisations, like Commission Junction and LinkShare, their offering is too great for affiliates and publishers to turn their back on. However, Google may take a huge hit to their profits to achieve market share, at which point networks may have a real battle for publishers and affiliates on their hands.

Digg This! | Save to del.icio.us | Reddit | Submit to Netscape | Google Bookmark This | Furl This | Add to Yahoo MyWeb 2 | Contact bigmouthmedia

Google wins Kinderstart lawsuit over PageRank manipulation claims

Mon, 26 Mar 2007 12:00:00 GMT

Google's increasing popularity and fame does have its drawbacks, it seems. During the past few months the search engine giant has been inflicted by a stream of lawsuits, mainly related to copyright infringement.

One of the most high profile in the barrage of legal cases held against Google comes from KinderStart.com, a search engine and directory focused on children and parenting information, who claimed last year that Google had intentionally dropped their PageRank to zero in an attempt to destroy the competition.

Kinderstart.com claimed:

"Google has interfered with KinderStart's First Amendment rights [to free speech] and has engaged in predatory conduct and anticompetitive conduct directed toward achieving the objective of controlling prices and/or destroying competition."

It is difficult to consider Kinderstart.com, a rather simplistic search engine in comparison to Google, as a threat in any way to the search engine giant. Judge Jeremy Fogel, who was in charge of this Court case in California, dismissed the case for the second time this week after KinderStart failed to prove the allegations, labelling them as "vague, ambiguous and probably false".

Kinderstart.com was launched in May 2000, and aimed to help web surfers find specific resources and information related to young children aged up to seven years old, offering advice on a variety of issues from adoption to child development, pregnancy to nutrition tips.

According to Yahoo!'s link command tool, Kinderstart.com possesses approximately 8,900 inbound links. This reflects its huge popularity; however the site's PageRank dropped drastically and still stands at 0 out of 10.

PageRank was developed at Stanford University by the Google founders, Sergey Brinn and Larry Page, while doing their PhD research project. The idea was inspired by the system applied to important academic papers, through which the relative importance of a paper is determined by the number of citations it receives from other academic papers. In this way, links to websites could be thought of as 'citations', and when a site has many 'votes' or incoming links its PageRank increases.

PageRank can be regarded as the value or relative importance that Google assigns to any particular page and it is intimately related to Google's rankings. Google tweaks the PageRank algorithm frequently and many companies may notice a slight drop or increase in their PageRank when these changes do take place. However, when a site's PageRank drops completely to zero, this is due to a penalty being enforced as a result of the employment of spamming techniques, such as link building or keyword stuffing.

After eleven months of intense legal wrangling, Kinderstart's lawsuit against Google has been repeatedly dismissed, and the child-focused site may now also be made liable for Google's legal fees.

Digg This! | Save to del.icio.us | Reddit | Submit to Netscape | Google Bookmark This | Furl This | Add to Yahoo MyWeb 2 | Contact bigmouthmedia

Google through the ages

Fri, 23 Mar 2007 12:00:00 GMT

Google have recently been tentatively experimenting with a newly structured homepage format. Changes may seem small but for us search enthusiasts - whose lives revolve around everything Google - it is interesting to see this evolution from the giants of our online world. Perhaps it's time now take a look at these incremental changes and cast our eyes back through Google's changing interface over the years...

Google's Recent Reshuffle

Google Operating Systems reports that Google is currently testing a new homepage interface. Google has undergone a test move for the navigation links on Google.com. The navigation links move to a position directly above the search box and directly under the Google logo, towards the top left of the page and parallel to the sign in link.

It is not the first time Google have made alterations to their classic homepage but this new test move could be the most daring to date. The links to Google's services were detached from the search box and moved in the top left corner. Because there's no visible connection between the links and the search box, this move has allowed Google to add non-search services: Gmail, Google Calendar and Google Docs. Gmail and other services such as Calendar and Docs and spreadsheets have now become available for general use. As a result these developed services are being promoted to a more obvious and prominent location.

So let's look back...

The first images of Google's homepage are now fascinating. The company has come a long way since December 2nd 1998 when it featured a plump logo, 'Staniforth Search', Company info, Google Logos and e-mail subscription. The 'Staniforth Search' facility reminds us of the company's humble beginnings in Staniforth University.

This look remained until April 22nd 1999 when the homepage was refined to a much more concise format with only three link options immediately available.

On Oct 1st 1999, Google launched the logo that has stood the test of time. Many small interface modifications were made but up to Jan 23rd 2002 Google kept things nice and simple. On this date, Google implemented four main tabs to the homepage: Web, Images, Groups and Directory.

On Sept 22nd 2002 Google made a big introduction - Google News. News was the fifth main tab on the homepage.

The next main tab 'Froogle' was added on the 25th March 2004 and Google again simplified the interface by removing the coloured tabs, creating the image that we are familiar with today.

We now have an additional 'More' option as Google really starts gaining momentum expanding its product offering and the site remains fairly consistent from this point taking us up to the present day.

Digg This! | Save to del.icio.us | Reddit | Submit to Netscape | Google Bookmark This | Furl This | Add to Yahoo MyWeb 2 | Contact bigmouthmedia

Google to increase artificial intelligence wisdom

Mon, 19 Mar 2007 12:00:00 GMT

The variables that determined Google's algorithms are still as secret and obscure as the Coca-Cola formula; but SEO professionals are still able to predict and detect significant changes in the way the world's leading search engine works.

A minor change in Google's algorithms can leave out many businesses which do not possess a strong online presence, or rely on a single optimisation strategy. Being prepared for future alterations to Google's algorithms is crucial if you want your website to succeed on the SERPs.

Google's founders, Sergey Brin and Larry Page, are both considered pioneers of artificial intelligence and, not surprisingly, there is a large focus at Google on making their algorithm more intelligent. Machine learning implies recording actions, studying user behaviour and drawing conclusions from it. Users who install Google's Toolbar are contributing to the artificial intelligence development in Google, as all the searches made from the toolbar are recorded and sent to Google for further examination, in order to establish patterns of user behaviour.

Google is now stepping up their investigation of user behaviour in order to determine the site's popularity. Essentially, for any given search term, Google will study the number of users clicking on a particular URL. If thousands more users click on one page above others, then Google will increase the popularity of that page for the relevant keyword or keyphrase.

In the past, the concept of site popularity was purely based on the number of incoming links, and Google's entire business model was born from the idea where links were the determining factor on the equation - but all this has changed in recent years.

Google also allows the personalisation of search results depending on geographic region; this is known as 'geotargeting'. This indicates Google's clear intention to move away from 'monolithic results', a term coined by Google-guru Matt Cutts, towards a more customised solution that will make the user's search experience more democratic. The user will benefit from increasing geotargeting results, but those businesses relying heavily on links will experience a drop in traffic.

With the evolution of search engine optimisation now pointing in the direction of customised results, users now have the power to make a site more popular - or risk destroying a site's rankings if they repeatedly ignore a particular page. In this new era, it's important to remember that, however well a site is optimised, the achievement of higher rankings on Google universally is becoming increasing difficult as geotargeted results become more common.

As the Romans would say, 'divide and conquer' - diversify your techniques and don't rely too much on a single driving force to your site, whether this is in the form of links or otherwise.

Digg This! | Save to del.icio.us | Reddit | Submit to Netscape | Google Bookmark This | Furl This | Add to Yahoo MyWeb 2 | Contact bigmouthmedia

IceRocket: the one-stop-shop for the blogging world?

Wed, 14 Mar 2007 12:00:00 GMT

As new search engines grace the web world at such a fast rate these days, it's hard for fledgling search engines to make an impression on the billions of web-surfers who ride the Internet waves - unless they do something a little different. IceRocket.com is a blog search engine that, as well as a search facility, provides many tools for bloggers and blog-readers. IceRocket allows users to search through fairly standard things like news and images, but also has a search option dedicated to MySpace and blogs, the main focus of their site.

Bloggers are well catered for by the site and it is clear that IceRocket has been designed with the common blogger in mind. After every search a drop down menu is provided at the top of the page, which can be used to subscribe to a feed of the search results in various online blog readers, and a "Share" menu allows a link to be posted to Facebook, Digg or Newsvine. IceRocket also provides a free RSS Builder that is great for bloggers who are just starting out and will allow them to syndicate their feeds in the standard format for distributing online news. A free 'Blog Tracker' is also available so users can see how popular their blog is at present and over time.

IceRocket also offer a whole host of other blog-related tools that bloggers and other interested parties can put to good use. The main blog search can be modified to search for blogs that link to your site, as well as author-specified topics, or - as they are called in the blogging world - 'tags'.

The Blog Tools section of the site reveals a whole host of other trend-analysing applications. One of the most interesting tools is the Blog Trend graph maker. With this tool users can specify a keyword, search blogs to see which keyword has been creating more buzz and tailor their topics as such.

More well-implemented features are the Top Videos, Top Movies and Top News pages, which IceRocket compiles by trawling blog-links to online video sharing sites, such as YouTube and Google Video for videos, IMDb for movies and major official news sources for news.

As IceRocket grows it should become even more useful as it trawls more blogs and its web presence becomes more established. The site looks to be heading in the right direction for bloggers and blog readers and by providing and adding to these tools they may soon become the 'one-stop-search-shop' for which many writers in the blog world are desperately looking.

Digg This! | Save to del.icio.us | Reddit | Submit to Netscape | Google Bookmark This | Furl This | Add to Yahoo MyWeb 2 | Contact bigmouthmedia