Published Jul 22, 2003

Matt Haughey has recently added his name to the row being raised by The Register suggesting that blogs be removed from search results at Google. Now, when an A-list blogger and longtime proponent of the medium suggests such a thing, well, it may be time to seriously consider the proposition.

The controversy centers around a simple set of observations:

  1. People use Google extensively to search for stuff on the Web
  2. The ability to find stuff is key to the usefulness of the Web
  3. Therefore, it’s important that Google continue to return high-quality results
  4. But, increasingly, blogs are showing up as top results in many searches
  5. This is because blogs are frequently-updated, heavily linked-to, and link, in turn, to other heavily linked-to sites
  6. Blogs also tend to have better-structured information that Google’s bots have an easier time handling and relating to other information
  7. As a consequence of blogs being rated so high in search results, sites selling items and news sites reviewing items receive comparatively low rankings in search results
  8. It is these news and sales sites that users are trying to find
  9. Blogs, therefore, obscure the sites that users are looking for

Writers at The Register have long held that blogs should be removed from Google results before they drown out all other results. But that position relies on assertion #8 above being true. I tend to disagree with that, and here’s why:

  • Searchers on the Web are looking for relatively specific information
  • This information involves specific items (concepts, etc.)
  • Searchers are looking for information on these items (concepts, etc.) that is relevant to them and their particular situation
  • Sales (advocacy, etc.) sites provide only information designed to turn the visitor into a customer
  • News and review sites provide information targeted at a wide demographic, or at a very specific (and usually clear) demographic
  • Bloggers provide information about their knowledge and experiences

So, what of the above information best matches the searcher’s needs? Well:

  • Sales sites only match if the searcher is planning to make a purchase and if the searcher is not going to make that purchase from a well-known and trusted store, such as amazon.com, bestbuy.com, etc.
  • News and review sites only match if the user is in their demographic
  • A blogger may have experienced exactly the same set of needs or events that the searcher has — they may have looked around for a new religion, or tried to buy a DVD burner, or whatever.

To me, it looks as if the blog result may actually be the most relevant of all listed. For instance, Matt Haughey talks about his TiVo blog turning up as a high result for TiVo searches. That’s exactly the site I’d be looking for, with information on usage and upgrading and personal experiences that might be like mine. Why drop it?

Now, blogs aren’t perfect. They tend to be incestuous, absorbed with minutae and even masturbatory. Worst of all, some lack any kind of peer review or outside responsibility (some bloggers, however, have made their name from providing high-quality information on their sites — and it’s these who are the most-linked and therefore would likely turn up highest in Google results). Some proportion of blogs are likely to be the worst results returned. It’s probably best to remove those results. But how do we do this without removing blogs in general? Here’s a few ideas:

  1. Relegate blogs to a different section of Google, and return no results in standard searches. This fails the smell test because it discards a substantial portion of useful results.
  2. Let the user selectively exclude blogs from search returns. This is a reasonable approach, although it begs the definition of a blog. It also would likely have to be relegated to Google’s advanced search options, given that company’s unwillingness to complicate their front page in the past.
  3. Maintain a higher standard of accountability for blogs. Google bases much of its ranking on the quality of incoming and outgoing links on a site. If blogs were required to have more (or better) links to achieve the same PageRank as a non-blog site, this would mitigate the excessive influence of most blogs, while still ensuring that the best bloggers got the results that they deserved — and that searchers were looking for. Again, this begs the definition of blogs and also is invisible to the user, who doesn’t know that certain sites are being rated comparatively poorly.

None of these approaches are perfect, but some combination of the three (in fact, all could be implemented) would increase the quality of Google’s search results without excluding a large pool of potentially useful sites.

1 Comment

Follow-up:
Ooh, somebody did real research into things and found that blogs didn’t negatively influence results:
http://microdoc-news.info/home/BloggerNews/2003/07/23.html/1