Comments on Four Modes of Seeking Information and How to Design for Them

Mary Deaton — Fri, 26 Jan 2007 14:52:43 GMT

History and bookmarks are supposed to be the browser’s way to let you refind, but I am sure many people are like me and do so much searching and have bookmarked so many things that I may not even remember having done it.

Perhaps the solution is for browsers to become smart enough to analyze search text and tell you if you already have matches in your history or bookmarks (or let you search your history or bookmarks!)

Any site that lets you store bookmarks should also do it. I use del.icio.us and have noticed that it lets me post the same page twice!

Google Desktop searches to see if there is anything on my computer that matches my search, but I don’t think it searches my bookmarks or history. Should it?

Ruben Timmerman — Fri, 26 Jan 2007 14:52:41 GMT

Nice article, love the examples. Have a few to add:

Maybe a good example for the exploratory users: a breadcrumb (which shows the hierarchy of navigation options). These users make more “mistakes” while browsing, and might be helped by quick overviews on other navigational levels.

Another nice semi-active-passive form for refinding users is Google’s account option that saves all your searches and the results you clicked on.

A passive solution for refinding might also be one of the simplest: indicating visited links. Too many sites forget to give visited links a distinct color.

I wrote about and expanded on this on my blog (for Dutchies, sorry!)

Daniel Szuc — Fri, 26 Jan 2007 14:52:41 GMT

Super read!

Much of the “3. Don’t know what you need to know” can be helped my developing product pages that are written in ways that the users understand. There are still so many product pages that are too technical in nature and dont assist a purchase process. Examples include: banking, insurance, computing, telecommunications product pages.

Andres Zapata — Fri, 26 Jan 2007 14:52:40 GMT

Google Desktop has a (big_brother_type) feature that keeps track of EVERYTHING you do on your computer – this can be sliced by whatever (mainly date) – very powerful. I thought of that while I was reading about “re-finding.” What can we learn from Google? (and no, I don’t work there) :)

Simon Smith — Fri, 26 Jan 2007 14:52:40 GMT

Enjoyed the Article, Though provoking:)

Would it be fair to say that the way search engines operate has affected the way that we search, making exploratory search more common, even for known item.

To take the example of the known item, if using a web search engine, my experience is that even if you know what you are searching for you enter one or two words related to your requirements, see what you get, refine the search terms slightly and so on. This is because entering a very precise query often either returns no results or the, engine is “confused” by other words in the string and returns many only loosely related items. Therefore you are using “exploratory” methods to find or re-find known items.

Regardless of how you seek the information, I think that the most effective way of meeting search requirements is a degree of personalisation. If you know who a person is in terms of what they have searched for previously, how they search and can derive an understanding of the intent of the query, but also understand the relationship between query and answer you can deliver a much better answer much quicker.

Scott McDaniel — Fri, 26 Jan 2007 14:52:40 GMT

Nice article! :)

My first impression was that re-finding might be a special case of of Known Item searching. What would you say are some of the searching approaches or UI requirements that would differentiate them?

For any re-finding task, a history of visited pages or previous searches would help.

You can also use the approaches to generate more specific design guidelines. See this article that my wife and I wrote for User Experience.

Leo Sauermann — Fri, 26 Jan 2007 14:52:40 GMT

Scientifically, it looks hard to differentiate between

* 1. Known-item and * 4. Re-finding

Why did you differentiate between them?

btw: the “finding and reminding” paper by bonnie and nardi is related.

Jenny Reiswig — Fri, 26 Jan 2007 14:52:39 GMT

I work part of my week at a library reference desk and I see these modes every day. But it’s a truism of reference that people never ask you their real question, so getting to their true mode often requires some negotiating. A user looking for a known item will often start out with a very broad question that looks exploratory – “do you have information about cancer” when s/he really has a specific article to find. When I start to answer the expressed question, I can see the mismatch on the user’s face… how do we see it in our systems?

Shuan Lo — Fri, 26 Jan 2007 14:52:39 GMT

Thank you for the interesting article. I’m new to the field, and these are new concepts to me.

After reading this article, I’m wondering if it is possible to classify information seeking behavior according to two facets: 1) precise-vague, 2) aware-unaware.

Specific meaning the user knows exactly what he or she wants to look for; such as tomorrow’s weather in Taipei. An example of vague could be “how is it like to live in Taipei”.

In terms of design parameters, maybe the precise-vague axis maps crudely to a search-browse axis.

I’m not sure what the aware-unaware axis can map to. Maybe a passive-active axis? A passive feature would be like my browser’s record of my history and Google’s Desktop Search. Passive in the sense that the user initiates the use of these to seek for information. An active feature would be like showing related news headlines beside a news article. Active in the sense that information is being pushed to the user.

Perhaps an active-search feature could be my Firefox’s Google Suggest extension, and an active-browse feature could be a list of “most viewed”?

I’m not sure if these make any sense.

Donna Maurer — Fri, 26 Jan 2007 14:52:39 GMT

Great example Eddie. So think about how to answer the questions they do know about, and show what you need them to learn.

C. Sanchez – refinding is a sometimes a known item task, but sometimes not. Eg on boxes and arrows, I often look for my own articles (so I know exactly what I am looking for and they are easy to find) but sometimes want the one that has info about page description diagrams (I go back to this a lot, but never remember the title or whether it was Dan Brown or someone else). A list of my most-visited articles would help in both examples. But yes, there is definitely overlap.

C. Sanchez — Fri, 26 Jan 2007 14:52:39 GMT

Good article.
I’m not too clear on one point however.
It seems to me that ‘Re-finding’ and ‘Known-Item’ modes are quite similar.
I’m not sure I see the distinction.

Eddie Cook — Fri, 26 Jan 2007 14:52:39 GMT

“Don’t know what you need to know” has another twist. I train technical support/ customer service. Quite often my company has a policy or solution that would be a perfect fit for the customer, but the agent isn’t aware it exists, and so, doesn’t begin to search for it. We use weekly reading time to get caught up on corporate headlines, but there’s always oodles that’s missed. I’m beginning to experiment with a hierarchal search that displays an index of sub-catagories to a product or system, similar in feel to Windows Explorer. My hope is to have it used as a roadmap for a process flow during a call. I would welcome any suggestions to make it effective.
Excellent article. Very thought provoking.

Andy Vaughan — Fri, 26 Jan 2007 14:52:39 GMT

Re-finding: You’re *definitely* not alone in this!

There have been countless times when my brain has taken in information from one of the many books or websites I’ve seen, conversations I’ve had, TV programs I’ve watched etc. but I didn’t make a concious decision at the time to log it somewhere so I could find it again. Then, 2 weeks, 2 months or 2 years later, I’ll find myself in a conversation about that topic or situation where I need the info but I can’t remember where I found it the first time round.

Can be really frustrating!

Donna Maurer — Fri, 26 Jan 2007 14:52:38 GMT

‘Don’t know’ is a little different in that they may think they want one thing, but something that they don’t yet know about is more important. Like the terms and conditions example – you want to sign up, but there might be something in the t&c that you just need to know because it is really important. But you don’t go on an exploratory journey to which terms and conditions is the end result.

Looking at the design solutions too, short, simple answers don’t necessarily satisfy an exploratory seek, but will satisfy a ‘don’t know’. Getting them to related material helps both.

Keep asking me to clarify – it helps me to think about how to make my fuzzy thoughts clearer!

Austin Govella — Mon, 30 Apr 2007 09:00:53 GMT

Donna, exploratory and don’t know… seem like very similar behaviors. Is the distinction between the two simply that exploratory has no specific intent where don’t know’s have a fuzzy intent? Or is there a brighter line you’re drawing?