The problem with Google Search Results: Why being “popular” and “web native” sucks

Google results are crap, but we use them anyways.

Many people are trying to game Google. In fact there’s a whole industry called Search Engine Optimization that focuses on doing this, many of them I know personally.

Here’s two reasons why Google Search results suck:

Google Results return most Popular
The problem with Google search results is that it returns the most “Popular” content, which doesn’t mean it’s always the “Right” content.

For example, Scoble often tells folks he’s “The number one Robert on Google”, yet there are Roberts that are more well known than him, such as Robert Redford or Robert Dinero.

Google yields content only on Web
This leads to problem two: Google delivers the most popular webpages or sites that exist on the internet, and it you’re not on the internet, do you matter? The problem is, Robert Redford and Robert Dinero have a stronger precence on the silver screen and TV than they do on the internet.

I come up top for Web Strategist, and Social Media Measurement, but does that mean I’m the best? the most accurate? nope, not at all.

So what will happen? How will we evolve? Boss John Furrier suggests that we look at social networks, communities, and those we trust to find information. I’ll bet part of the answer is there.

  • Over time I think you’ll see Google get away from returning purely the most “popular” results – especially if you’re signed into personalized Google accounts. They’ve already made some improvements with universal search, but that depth hasn’t made it to the majority of search phrases. In the future, when someone types in Robert, you may still see Scoble, but you’ll also see video (maybe a Dinero or Redford Interview) and news (the most popular Robert of the day perhaps). Search as an industry is about to enter into a stage of rapid development with personalization and behavioral targeting.

  • I agree with Pete, to some extent. I think Google will eventually start to move away from “popular,” at least somewhat.

    Unfortunately, at this stage in the game, they use “popular” as a way of determining relevance. If a lot of people are linking to something, and using certain anchor text to describe what they’re linking to, then Google is going to assume that the page in question is highly relevant for that term.

    Of course, in reality, everyone knows that popularity and relevance aren’t necessarily the same thing. If they were, there’d be no such thing as a Google Bomb.

    But, I do think that, eventually, Google will move away from a strictly popularity-based algorithm (in fact, I think they’re already starting to do this). I can’t personally think of a more effective way to determine relevance in search, but I’m sure someone is working on it. If not at Google, then somewhere else.

  • Pete, Adam

    Good points about Google IG as a data point to refine the search results.

    Do you think social networks will impact this and help aid information discovery and delivery?

  • I definitely suspect that users will start to use social networks as a way of getting information that is more relevant and specific to their needs.

    How or if this will affect search engines, I’m not sure. They may start to place more weight on links from social networks (though, the popular opinion in the SEO community is that the exact opposite will happen very soon, so it’s hard to say for sure).

    Regardless of what the search engines do, I certainly agree with the suggestion that we turn to social networks to get relevant and specific information from those we trust.

  • Results are only as good as the search query though.

    Searching for a term as generic as “Robert” or “bike race” isn’t likely to help anyone find specific information about Dinero or the Tour de France and in the case of those more generic queries, frankly, I’d prefer Popular — because I view those results as more relevant (given the framework of the query).

    So in this case, it’s not Google but rather the query writer that sucks.

    Google continues to experiment with post-search navigation and I expect we’ll see significant improvement that will assist people in their search for relevant and popular information on generic queries (remember when the Northern Light search engine was online?).

  • Chris Collins

    Cant say I disagree with your comment about gogole search results, but I would over gaming the system.

    Certainly some people are trying to “game”, however there are plenty of people who arent even “in the game”. They simply have a website that isnt putting its best foot forward.

    En example of gaming to me is the case where I go and repeatedly have url’s pointing to my website on peripheral website[s] (as many as possible), to effect the page rank algorithm. We saw this with things such as “Whitehouse.com” a few years back. Since then, google has got smarter about such obvious gaming techniques.

    When it comes to SEO and at least the scientific part of SEM I am not sure they are trying to teach you gaming, its really is optimization of your content for the killer app “search”. If you don’t point out what your website is about you will never be found. For instance, if your a plumber and you don’t mention that you offer plumbing services in the first few paragraphs, in bold text or in the title of the web page would you identify that page as being about plumbing? In general the textual part of ranking a page for similarity takes into account some general document observations. These are, words that you use in a search that are rare in the language must be more important than more commoon words (scoble is a rare word compared to “computer”). If that word is in some kind of highlighting tags like Scoble then the topic of the text is very likely defined by that title and its deemed even more important….these are generally the types of observations and techniques brought out in SEO plus a whole lot more.

    Secondly the definition of popular. Well seems that Robert it probably the most popular Robert “on the internet”. Ranking algorithms take account of how popular a website is (page rank or a more trivial in link count), how relevant the words used in the search that intersect with that web page and also the amount of times someone clicked on that page in search results….yes they will learn that if people don’t click on the result then it must be relevant. This feedback loop takes care of documents that look good from a pure ranking function perspective, but in reality suck.

    In all ranking is subjective. People have something in their head about what is relevant and a computer algorithm isnt a mind reader. What it has at hand is popularity based upon prior click examples and its relative page importants compared to the other 8 billion + pages.

    p.s. I build search engines for a living and I am not a SEM/ SEO consultant.

  • Pingback: lauren’s library blog » Blog Archive » links for 2007-07-14()

  • It points to the potential for the role of personalized search and social search. Thanks to iGoogle, personalized search is now becoming more prominent, but social search remains a muddled mess. That can change in a flash though.

  • So what do you think about Neilsen’s switch to ranking sites based on time spent?

  • Hiya Lisa! Are you going to ISF this week in Portland?

    Regarding Nielsen, this was a smart move, but it’s still somewhat limited.

    Attention is just ONE attribute to be measured. It has weaknesses as users can leave a browser open with tabs all day and night, and that wouldn’t be a good measure

    We’ve talked about this in depth here:
    http://www.web-strategist.com/blog/2007/07/13/the-conversations-about-social-media-measurement-grows/

  • Pingback: BuzzerMatrix-Open Buzzer Blog » The problem with Google Search Results: Why being popular and web native sucks()

  • Pingback: trust, word of mouth and the conversation « ditigal ramblings()

  • Pingback: Web Strategy Reading: Oct 28th, 2007()

  • The phrase you are looking for is:

    Give us ‘relevance’ not ‘popularity’

    Anyone can get what looks like popularity by trickery. Popularity is scored too high. It may be a factor, but smaller than currently scored.

    Content is king? That’s what they used to say.
    Bring back ‘Relevant Content’ as king!