• http://www.highrankings.com/newsletter/ Jill Whalen

    Interesting theory.

    What do you think about negative references? What if a company sees a surge in people saying that they suck, for instance?

    Do you think Google would know the difference and not count those as something good, or even more so, count them negatively towards the company in question?

  • Andrew Goodman

    Just riffing here, but it’s maybe worth going back to the whole concept of peer citation – from which this experimental web-based version (HITS, PageRank, etc.) grew – and asking just what the heck was *it* supposed to accomplish in the first place?

    Garden variety academic citation analysis:

    * Made no distinction between positive and negative mentions

    * Generally made no distinction between “top” journals and books, and second-rate ones

    * Counted the total number of citations

    * Only counted publishers and journals that were on the radar (peer-reviewed, accredited publishing houses, sometimes a major periodical); did not count other mentions.

    What would a scholar get for being well-cited?

    If they were an absolute superstar: probably bargaining power in the marketplace, and promotion to a University Research Chair or some similar superstar position with a high salary and little or no teaching.

    Advanced citation analysis might have worked a bit more like PageRank, in that you also count the authority score of those doing the citing.

    In the end, though, no big prizes for being well cited. Everyone gets to work in their career and there is no winner-take-all economy.

    For ranking well in SERP’s on the other hand, the rewards can be substantial. But there is no real gatekeeping process of any kind, just this free-for-all of different kinds of actors (millions of them) free to exhibit “recommending” behavior.

    Chaotic, frequently gamed behavior resulting in sometimes huge rewards for the winners.

    Which sometimes makes you wonder.

    It’s an interesting, emerging science, but in spite of all the data coming from real people out there and despite the use of human “SERP satisfaction” raters, most of us don’t get a vote on the principles the search engines use to decide rank order.

    We as an industry just sort of bumble along, assuming there are good reasons for everything.

    Eric offers the wise understatement of the year in his point that “the wisdom of the mob is inherently quirky.” And not only can social networks be elitist, we want them to be (and we will perhaps continue to rely on that customized elitism will return better recommendations to us from peers). Pushing someone else’s (everyone else’s) peer system across to the general search results seems like a recipe for misusing those signals.

    The problems are not completely intractable, but it’s high time our industry started having intelligent conversations about the principles behind information retrieval. I think we frankly did more of that back in the day, 10-15 years ago. And why not – there were far fewer dollars to chase back then, and a lot less data and signals to drown in.

  • http://www.ericward.com Eric Ward
  • http://www.ericward.com Eric Ward

    Sorry about that, my a href tag broke the post…Trying again below.
    I remember making an off-hand remark during an SMX session in NY in 08 or 07 that links don’t have to be links. That URLs without an a href=”” tag which were not even clickable could in fact be a helpful signal, or citation. And it makes sense as you describe. Imagine someone with true subject expertise mentioning a site’s domain, in print and online, but without it being coded so as to be clickable. How is that any less valuable? Answer: It isn’t, and could even be more useful than a clickable link, depending on the author. And as acadmic journals migrate more and more to the web, many archives come along with URLs that aren’t clickable links either. Dig through Google Scholar and you’ll find plenty. It’s the presence of a URL on a page of merit that matters, whether clickable or not.

  • http://www.seopros.org Webmaster T

    Eric, citations have likely been used in content discovery for some time based on the patents Google has and the fact Google basically will use any app or mention in it’s quest to index/discover every page on the web. There have been discussions of this among SEO’s for quite some time. I have used a citation search rather than link search on Google for over six years because IMO, it has always provided more info then the link: operator on either Yahoo! or Google. However, that said I do not see any indication these are suddenly more valuable… IMO most people just didn’t bother looking for them. Jill is right about the discovery of negative stuff and potentially people selling info they don’t want you to find or doing things they don’t want you to find during analysis of IBLs.

  • http://glager glager

    How would search engines handle more ambiguous ‘brands’ such as people’s names? Would all the hype around Kanye West right now get his site to rank higher or his wikipedia entry? Which is more relevant to “Kanye West?” Would a buzz around Danny Sullivan help to improve the rankings Search Engine Land, Sphinn, SMX, or Daggle?

    I agree that this is the way things are headed but it seems there are very difficult problems to overcome at this point.