SearchMash: Google Reborn

Improved Search Algorithm

SearchMash may have started out as Google with a makeover, but now it’s radically different on the inside. Whereas SearchMash used to have identical results as your normal old Google, the results list now varies not only in order & ranking, but even in the links available.

SearchMash doesn’t use its own page index – that would be taking reinventing the wheel to a whole new level. But SearchMash does analyze the contents (and especially the relevance) of websites differently; bringing links to the first page of results that would have been missed entirely otherwise. To date, no one really knows how Google’s search engine algorithm works (obviously), except that it prioritizes back-links over content most of the time.

The first glance at the SearchMash search results are much of the same. For instance, the infamous “miserable failure” search brings up the same result as Google does. But that’s just the tip of the ice-berg. One of Google’s shortcomings, the higher rankings of news aggregation sites, is a direct result of backlink prioritization. When sites like Slashdot and Digg link to a relatively small website, it is not uncommon for the aggregated links to rank higher on Google than the actual article itself1.

In our subjective testing (seeing as there doesn’t exist a truly random way to search Google and SearchMash for links to small sites that appeared on Slashdot and Digg), SearchMash seems to do better. It appears that SearchMash has either defined the list of user-submitted news aggregation sites (and the URI patterns) or (hopefully!) that it dynamically analyzes the content of such pages for excerpts from linked pages. Either way, what this means is, no more less Google bloat.

Less Google Spam. That may sound impossible, but whether its placebo or the real deal, it seems to us that SearchMash has a bit more relevant content and that much less spam listings. Especially when we tried to search for the traditionally much-spammed “media converter” topics (and its variations), we saw fewer “media-converter.myspamdomain.com” listings and more truly Google-worthy listings. Again though, it might just be placebo, but that’s our take.

Honestly though, it’s too early to tell. SearchMash has only been out for 2 months or so, and until a couple of weeks ago, its results were identical with Google’s. Even today, the difference isn’t that dramatic. All of the above was slightly exaggerated: basically forecasting the graph. SearchMash seems to be the beginning of something new and totally overhauled as far the algorithm goes, but only time can tell whether the improved results are a) placebo b) coincidence or c) here to stay. In the end, we can only hope, and revisit this topic at a later time to find out.


  1. This is in sharp contrast to MSN Search, which doesn’t give too much priority to backlinks and is more fair to the little guys, preferring content over popularity. Who’s right? Let the current standings speak for themselves. 

  • Similar Posts

    Craving more? Here are some posts a vector similarity search turns up as being relevant or similar from our catalog you might also enjoy.
    1. On the growing, intentional uselessness of Google search results
    2. Internet Explorer 7 RC1 Review & Open Search
    3. Seraphim Proudleduck Revisited
    4. 5 Things Yahoo! Does Best & Why It's Not Enough
    5. Google vs. MSN: What to Expect..
  • 7 thoughts on “SearchMash: Google Reborn

    1. I just looked at SearchMash.com after seeing a link to this story on Digg, you’re right, it does look like someone barfed the design up. But I really like the content blocks on the side, I think they’re something the rest of the search engines would do well to copy. It does bring context to otherwise overlooked blog entries, videos, images, and of course, our beloved Wikipedia.

      I actually enjoyed reading this review, loved ‘The (Inevitable) Conclusion,’ that’s so true! Anyway, nice review, I’m going to be keeping an eye on SearchMash for sure now! 🙂

    2. If you are down serveral pages in your search list and then follow a link that you are interested in, when you hit the back button to return to the list you have to start from the top of the list again to scroll down to the location you were previously at.

       

      Very annoying 🙁

       

    3. Hi Larry,

      Try pressing the <Backspace> button on your Keyboard instead in Internet Explorer. And in Firefox, hold down <Control> and then scroll the mouse one step back.

    4. I’m on IE7 too (albeit on Windows Vista), and ctrl+scroll, the back button, and the backspace key all work perfectly and take me back to my exact location on the SearchMash results page (or any other page for that matter).

      I didn’t notice that 100-result limit, nice catch Larry! Then again, I never looked more than 3-pages deep on any given Google search – so I don’t know if anyone actually looks that far down.

      At any rate, I can’t see any possible workaround, given as there are no GET variables (or even POST for that matter!) and the JavaScript function

      javascript:void(0);

      Does nothing unless a link is clicked :@

    Leave a Reply

    Your email address will not be published. Required fields are marked *