Fountain of Knowledge: IBM’s WebFountain

Years ago, out of IBM’s Almaden research campus, developers, analysts and computer science geniuses birthed IBM’s WebFountain.  The WebFountain project has been touted as both a potential Google killer and a tool for high level data mining.  The true definition of the application seems to be a mix of both;   like Google’s search engine, the world’s online data is indexed and stored. Then, WebFountain allows the indexed data to be transformed into business trends.  

This type of analysis of online data is not a part of Google’s current offerings. WebFountain looks at the businesses goals, resources and that status of various projects, then processes and analyzes billions of documents and terabytes of information by using an efficient and scalable software and hardware architecture to help analysts find the jewels amongst the vast amounts of online data.

The WebFountain project is a significant investment for IBM, which at one point had 200 research engineers and hardware resources capable of storing 167 terabytes of compressed data assigned to the project’s development.  The insights produced by this application are much more intricate than the needs of most online searches – or my most recent Google search, namely my attempt to find at-home recipes to mimic the menu at Roscoe’s Chicken and Waffles.  
In one successful example, WebFountain was able to help oil conglomerate British Petroleum (BP) discern that, at the pump, customers were concerned about crime and safety.  Through online data mining BP started offering free coffee and doughnuts to attract police officers to certain stations.  There are a few jokes to be had in the cops and doughnuts marketing strategy, but also increased safety at the pump.  

While the WebFountain project has had “stumbles” in development and growth, I thought it was a very interesting application.  This may be a look at the possible future of online search and analytics, but due to slowness of the system, and the loss of key clients, the next iteration of a project like this may not come from IBM.  Nonetheless, developing trends and discerning correlations between consumer interests and the indexable data online is what’s next in search.  I should mention that the prototype for the WebFountain application was developed by Tom Costello.  Costello is an ex-Googler and his newest venture, online search engine Cuil.com, is making news and directly challenging Google.com.

Article by Joel Collymore

 

What is a monopoly?

As someone who wondered why it took the government over a year to approve the Sirius / XM merger, I now wonder what is going on with the Google / Yahoo! advertising arrangement.  Based on the share of searches, Google would control about 90% of the search traffic.  I am no math major, but I swear that 90% constitutes a monopoly.  

This partnership is written as “non-exclusive” – which means that Yahoo! could work with any search partner.  It seems as if the language is there just as a protection against the government for 2 reasons.   First, how many search partners are out there?  Yahoo! already has turned down Microsoft and a deal with another search provider would do nothing.  Second, why would Yahoo! work with another search partner when they expect to make $250 - $450 million in the first year?  

The Sirius / XM merger is not a true monopoly because you don’t need satellite radio… I mean, I do and I don’t understand people that don’t have it, but it’s a choice.  However, what marketer doesn’t need search?  Where else can a marketer go to mimic the performance they get from search?  They will now be operating in a world where one entity owns about 90% of the inventory, and whose pricing model is highly secretive. To further complicate this for marketers, the deal has been rumored to increase PPC rates on Yahoo! by about 22% as well as some language that this deal will also allow Google to sell ads all over the web.  While AA|RF is owned by Microsoft, we are not privy to any information as it pertains to Microsoft’s proposed takeover of Yahoo!.  What we do know is that if this deal is approved, the search world just got really, really small.

Article by Joshua Palau

 

Strategic Link Building – Part 2

This week’s installment provides context around the mechanics of link building, as well as defining link tiers.  Read the first installment of Strategic Link Building from SMTrends 108.

The Mechanics of Link Building
As much as link development relies on relationships, analytical process is necessary to define a site’s link building objectives. Useful strategies can be gleaned from analysis of site conditions and appraisal of target site and competitor backlinks.

At the most basic, an ideal page for link placement will have relatively high trust metrics (PageRank or crawl frequency) and a low number of outbound links. This ensures that page not only has value to search engines but also will provide a significant boost to the target page. A page must also not contain “noindex” or “nofollow” tags that may affect outbound links, and its domain must either have no Robots.txt file associated with it or, at least, a Robots.txt file that does not place restrictions on the site section where the target link may be included. The best placement for the target link is within relevant body copy, and ideally should contain strong keyword anchor text. These are, of course, rather simplified qualifications for a link, but they warrant mention as a proper starting point in a discussion of advanced link analysis.

Avenue A | Razorfish assesses links according to a loose system of link tiers. These link tiers span the timeline from site establishment to maturation and include:

•    Formational links
•    Brokered links
•    Community links
•    Authoritative links
•    Naturally generated links


These link tiers cover the terrain of directories, link advertising, blogs, social media, relationship-based linking, and naturally occurring links. One may notice some tier overlap at times due to the integrated nature of the online landscape, and strategies can and should be designed to address this. A website must be evaluated as to where it fits into the spectrum in order to determine the starting point for strategy. This helps determine what types of links will aid natural backlink growth and what ones will hinder or have little effect based on the conditions of the site.

Link Tiers Defined
The following are explanations of Avenue A | Razorfish’s five link classifications. They range in order from formational to naturally generated links and encompass the maturation process of a website. A website continuously acquires links throughout its lifespan and needs to secure different types of links as it grows. Link tiers indicate the next level of links needed to thrive in the maturation process in regard to authority and link popularity. No link tier should be viewed as separate from or greater than any other classification, as tactics tend to overlap and aid websites in different ways.

Formational Links are the initial set of links one should pursue in establishing a new website. These links are sometimes overlooked during site launch and can be addressed up until the point that a website achieves a moderate level of backlinks or measured trust. This is subjective to the site and must be judged by the type and quality of backlinks observed. Formational links include links from directories, easily placed links and relevant niche links.

Brokered Links are secured through negotiation or agreement and can often produce moderate to high authority links. Brokered links often come about as a result of direct analysis of opportunities found through backlink research. They can involve a good deal of individual negotiation and must be addressed on a site by site basis, which may be time consuming. Many brokered link opportunities are bartered arrangements between sites that, if done correctly, can elevate a site out of obscurity.

The more controversial side of the brokered link tier is link buying. Several services offer bulk link placement as a method to increase the number of search-friendly backlinks for specific keywords. Google does penalize for explicit link buys, and the risk associated with this tactic is significant. If one is looking to attempt link buying, it is best to do so sparingly and slowly. Link buys boost search visibility in the short-term but will require the use of other tactics to sustain ranking.

Community Links reflect much of the Web 2.0 landscape, referring to blogs, social media, and user generated media. These links may be gained by engaging participants in the community on their terms through link bait, relevant content and social opportunity. These links can snowball if done correctly, generating a nature stream of backlinks. Tactics such as link baiting may result in a traffic spike that raises the search visibility of a page, but direct community involvement can foster long-term benefits.

Authoritative Links are high value inbound links from trusted online sources. These sites have high trust metrics and transfer the most link popularity per link. Authoritative sites include trusted news sites, .edu pages, and other sites regarded highly by search engines. A site should examine its backlinks and existing offline relationships to determine any opportunity for high quality linking. While these links are the most likely to remain active, the one downside of authoritative links is that they commonly link to the domain of the target site and not individual pages. The effects of this may be balanced by savvy internal linking.

Naturally Generated Links refers to any links secured without direct pursuit. In reference to link building, these links can be news articles referring to the target site, scrapers of quality “nofollow” sites such as Wikipedia, links from press release circulation and links for sites monitoring RSS feeds. Naturally generated links are the end goal of a link development campaign as they reflect continued link stability. A healthy influx of naturally generated links will help keep a website fresh and competitive indefinitely. Link seeding strategies are needed to make this link tier effective for client campaigns.

Article by Ron Sansone

 

SES San Jose 2008

Next week, SES San Jose kicks off and we will have several members of our search team serving as panelists.

On day 2, Analytics Engineer, Brian Cosgrove will discuss how to Identify, Analyze and Act: SEM by the Numbers.

On Day 3 SEO and Web Analytics VP, William Flaiz will discuss Building a Search Friendly Site, while VP & Global Discipline Lead of Search Marketing, Matt Greitzer will focus on Auditing Paid Listings and Click Fraud.

 

 

Subscribe / Unsubscribe                                  Privacy Policy                         About Avenue A | Razorfish™

The editor of SMTrends is Joshua Palau.  Please send any questions, comments or topic suggestions to smtrends@avenuea-razorfish.com or mail to: SMTrends, 417 N. 8th Street, 2nd Floor, Philadelphia, PA 19123. For more information, you can call us toll free at 866-858-1993 or email us at info@avenuea-razorfish.com  2007 Avenue A | Razorfish (TM).

 
Permalink