louisrosenfeld.com logotype

Home > Bloug Archive

Mar 26, 2002: Search System Trick #71

Here's an interesting idea: for each search result, show other searches that have retrieved this particular document. (Obviously, these other searches would have to be displayed on a separate page linked from the result itself.) So instead of "find other documents like this," we'd "find how others found this document."

Do you think it would be useful to see how other people found what you found? You could then actually execute one or more of their searches to see if it did a better job than your own search. Has anyone seen anything like this before?

If this is a brilliant new idea, all credit goes to Caleb Rutan of TechStreet.

email this entry

Comment: Mathew (Mar 26, 2002)

I think it could work well to get you thinking about other search terms you hadn't thought to include. A serendipity search.

I haven't seen anything specifically like that for search terms, other than the usual "popular searches" shenanigans.

Comment: Madonnalisa (Mar 26, 2002)

My professor(Michael Twidale) at UIUC-GSLIS was working on a prototype for the UIUC library systems for collaborative searching. I believe the system was called Ariadne:

Comment: richard wiggins (Mar 26, 2002)

When I get naive email from users of our search engine, and when such queries are trivially solved by a straightforward search, I email back the exact steps of the successful search.

For more complex searches, it could be even more useful.

So: it's a great idea, but no one reads tips. :-)

Comment: Paul Nattress (Mar 27, 2002)

As an IA, I would find a "find how others found this document" feature very interesting. However, I think users are far more interested in the actual results than the routes to the results. (This feature could be coded into the search function but only made visible to the developers.)

I would prefer to see a "other users who searched for XYZ also searched for ABC". This could help users who aren't sure what the correct search term would be for a particular subject. Using this system, they are more likely to stumble across the correct search terminology by following the trial and error paths of other users.

Comment: Eric Scheid (Mar 27, 2002)

I have the Yaywastaken referrer tracking system running on the IAwiki, and although there are *many* referrers from google & yahoo they are all pretty homogenous. Very little to suprise anyone.


Comment: carrie (Mar 27, 2002)

My friend's recently completed master's thesis - an online database of a French graphic arts magazine - has a keyword search. The results are a list of similar keywords with a total number of items for each, with links to those actual items.

For example, a search on 'photo' would give back 'fashion photography' and 'photogravure'.

It's a nice way of narrowing your search without telling the user *how* to search.

Comment: carrie (Mar 27, 2002)

the web site mentioned above is here:


I put it in the 'URL' field and it did not display with my comment!

Comment: Caleb Rutan (Mar 27, 2002)

Thanks for the mention here Lou, I am most appreciative.

The idea was spawned from a discussion which involved a number of people, Lou and Rich included.

There are a number of difficult aspects to this idea on an implementation level, and I certainly don't expect that it will help the average user. I expect it
will help people who search repeatedly, or who are attempting to narrow a very broad search.

From that aspect, I think the biggest questions I have on this particular subject are display oriented. Namely, do you display other methods of getting a particular result, or do you take a general sampling of the results, perhaps a cross-index of keywords they share and show searches that would have found that? I'm not sure how clear that is, might be simpler to show it as an example. If there is interest, I'll do that.

Comment: Keith Instone (Mar 28, 2002)

Attribute breadcrumbs are the same idea but for browsing: show the ways that you COULD have gotten here, not necessarily the way you actually got here.

My guess is that (all) attribute breacrumbs are based on hard-coded meta information, not on actual user behavior - that is a difference.

Comment: Luis Gomez Gonzalez (Mar 28, 2002)

Very Interesting, actually I had found better results while I do the research with a collegue, We think differently, So he has his way and I have mine, We put the efforts togheter and voila! This can be used for Web search or Document search.

It Can sure be usefull, People loose a good amount of time searching for information.


Comment: Eric Scheid (Apr 4, 2002)

Google shows another side of this same idea - for any given search term, show all other search queries which includes that search term.

The advantage here is that you can identify where your search term overlaps into another knowledge domain, and determine what keywords to use to exclude those peripheral results.

Add a Comment:



URL (optional, but must include http://)

Required: Name, email, and comment.
Want to mention a linked URL? Include http:// before the address.
Want to include bold or italics? Sorry; just use *asterisks* instead.

DAYENU ); } else { // so comments are closed on this entry... print(<<< I_SAID_DAYENU
Comments are now closed for this entry.

Comment spam has forced me to close comment functionality for older entries. However, if you have something vital to add concerning this entry (or its associated comments), please email your sage insights to me (lou [at] louisrosenfeld dot com). I'll make sure your comments are added to the conversation. Sorry for the inconvenience.