Why your web search results and mine are not the same

I recently asked, on the MHSLA listserv, for some help with finding a web-based resource about which I had very little information. I turned up a few possibilities; my colleagues, also running web searches, directed me to other sites that I had not found. While it’s likely that some of the difference in what we found lies in the varying approaches each searcher may take to answering a question, Marydee Ojala’s presentation “Searching, Finding and the Information Professional” (Computers in Libraries 2007 Powerpoint presentation) suggests five reasons that, even when we do key in on the same search terms, we may see very different results.

  1. Different search engines work from different databases. While 64% of all web searches are done in Google, searchers who use Ask (4%) or MSN (8%) or Yahoo (22%) will see results that don’t show up at the top of the Google results list, if they show up at all. (Statistics from Hitwise. I found them in Google.)
  2. Web search engines track us, and learn about us as we use them. They know where we are, and try to find materials that match our search history. Consider this: when I go to Amazon.com at home, the first screen recommends anime and science fiction titles. At work, it recommends health and hospital titles. Even when I’m not logged in! Web search engines also skew the results based on past use at a particular location. Imagine what this does to the searches at the library’s public-access workstations.
  3. “Squishy Boolean.” This concept, described by Mary Ellen Bates in 2005, includes the use of “preferred but not required” terms in addition to the standard AND/OR/NOT logic. Bates also describes some search engines that allow weighting of terms to indicate how important a particular term might be. I don’t see an obvious way to use “squishy Boolean” in any of the “big 4” search engines (Google, Yahoo, MSN, Ask) right now, but Bates wrote that she’d seen it in beta. (Bates ME. Squishy Boolean. Online 2005 Mar-Apr; 29(2): 64. Michigan readers, use this link to find the article in Gale.)
  4. Algorithms taking over. Sometimes, the search engines just ignore our carefully structured advanced-technique Boolean, and run the search against their default algorithms instead. I have seen this myself on a recent Ask search; a term I had specified should not be present turned up in the first five hits.
  5. Search Engine Optimization (SEO). Content providers use a number of techniques to increase the likelihood that their sites will show up when someone searches a related topic. Some web information services even “game the system” to have their results show up first. Ojala suggests a way to minimize the effect of SEO on how we see our results: use the “preferences” settings in the search engines we use to set the “number of results per page” as high as possible; it’s faster to scan 100 results on one page than to keep turning the page for the next ten … next ten … next ten …

3 Responses

  1. There’s no question that search results differ from search engine to search engine. There have been a number of studies that show that the overlap in search results from the different engines is a lot lower than many people think.

    At the same time, Google’s quality is not as steadfast as one might think. There also are studies that show that much of the perceived quality of Google’s search results is exactly that: perception. Without knowing what search engine is the source of the search results, many users rate results from other search engines higher.

    One way you can get quick access to search results from a wide variety of search engines is to use Zuula. You can get results from Google, but you also can get results quickly and easily from Yahoo and MSN … and from a wide range of lesser known search engines, such as Gigablast and Exalead.

    Good luck with your searching!

  2. Alex, thanks for the suggestion. I like the way Zuula presents results – tabs for each search engine make it easy to navigate, and “sponsored sites” are clearly distinguished. It includes some search engines I wouldn’t otherwise have tried, and the search techniques that I commonly use (+ to require a term to be present, quotation marks for phrase searching) actually worked well.

    When I first started searching, I used Dogpile, but I got out of the habit. Dogpile had some of the same infelicities as modern federated search engines; that is, since the source search engines don’t use the same conventions, the search results can range from “right on target” to “missed the mark completely.” I see Dogpile is still there, and works much better than it did back in the ’90s. The results are presented as a single “de-duped” list, which would save some time. On the other hand, results from “Ads by Google” and “Ads by Yahoo” featured rather prominently in my sample search results, just indistinguishably listed among the rest.

  3. You’re absolutely right, Sandy, about the key weakness of many metasearch engines. As you point out, most of them make it hard to distinguish “sponsored” results from true “organic” results.

    That’s why Zuula is refreshing in that it doesn’t do that. And, since I often find myself doing difficult searches, I’ve basically switched to using Zuula. I have rearranged the tabs for each search type, using the fact that you can “drag and drop” them. So I have my favorite engine for each search type as the first tab, and my favorite “backups” are next in line.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: