You don't know what you don't know

What do most people do to find information.  They use a search box.  The search box seems simple, elegant, and easy to use.  When you realize you need something, just go to the right search box, type what you want, and you have all the “information at your fingertips.”  It’s like magic.  So what’s all this about search being broken?

The first issue with search is that you don’t know what you don’t know.

Search technologies assume that you know when to use them.  In order to get the most out of them, you need to be able to recognize when there’s information out there that could help you.  Either you know that the information exists and you just need help locating it, or you think it might exist and make an explicit decision to stop what you’re doing to go and check.  That leaves a lot on the table.  There are too many times when, for one reason or another, you don’t decide to go look, and so you simply don’t know what you’re missing.

How many times have you completed a presentation only to find out that someone else in your company had already done one that you could have used?  Or maybe you’ve bought a product only to find out later that it has serious flaws, and when complaining to your friend, they say, “Oh yeah, I bought one too, I hated it.  In fact, it was so bad I blogged about it.  Didn’t you read my blog?”

While writing this entry, Watson found me blogs about Yahoo! search and search in Microsoft Office 12.  I wouldn’t have ever thought to go look for them.

If you didn’t know to ask in the first place how can all of the information now available help you?  Search engines can have all of the information in the world indexed and organized, available for you to question using powerful algorithms, but if you don’t think to stop and go find it, the information in them has no value. Making information available just doesn’t cut it anymore.  The information has to be made useful to you.

Like Don Norman suggested in one of his last essays, the simplicity of search user interfaces is deceiving.  I agree with this statement, but for different reasons.  The simple “elegance” of search user interfaces is a chimera.  Maybe it’s elegant in principle.  But in practice, if you don’t know what’s in there, how do you know what to ask for and when?

This Scientific American article really drives this point home.  The article goes into detail on current and future search technologies but stops with search as something that users have to go and do.  The future is a search interface for analyzing and visualizing richer data.  While this will obviously be useful for some tasks, it still won’t help you know what to ask for. 

You see, to me, search is something you do only because you have to.  It’s a last resort.  Search isn’t a task in and of itself.  It’s always instrumental to achieving some other goal.  You search because you’re trying to contact a company and you need their phone number.  You search because you’re looking for a product to buy.  You don’t go ‘searching’ for its own sake. 

If that’s the case then search should be tied more closely with what you’re doing.  Watson is our first step in making that happen.

Maybe the issue is the term “search.”  If what we mean by search is what we know today then, fine.  But if we take it to mean something broader—helping people get to the information they need---then today’s search user experience is due for an upgrade.

Apparently Andrei Broder concurs.  Who knew?  Watson did.  And it told me so while I was writing this.

But then again, none of this is anything all that new.  The difference is, now it’s real.

Comments

[MalkEvange]

Interesting article!
Where can I find more on this theme?