Login

Factual Blog /

A Good Answer is Hard to Find

Our goal at Factual is to build a platform that enables communities to transform bits and pieces of knowledge into clean, usable databases that can be leveraged across a whole range of applications.  But even as we make good progress on that particular front, I think about the other challenges that stand in the way of providing better answers to people’s questions, and the many innovative companies that are involved.

We all want good, reasonably accurate answers to our questions.  And, we want them to be simple and straightforward as well.  But unless you really understand the context, many types of questions just can’t be answered in a way that’s both correct and simple.

Conventional wisdom tells us that it’s easier to process information that’s been dumbed down, summarized, simplified. Because there’s so much information on the web, people will often times accept an imperfect answer if it also happens to be a simple one: Eat less meat…If you can buy, don’t rent…Get a degree from X University…See this movie in 3D…She’s just not that into you! It’s tempting to put faith in the soundbytes and knowledge chunks of modern processed media.  After all, who has the time to be an expert in everything?

But wait – this doesn’t sound quite right.  We all ask questions and often get really helpful answers.  Should I see this movie?  Where should I take her?  Is this iPhone app any good? Many times our friends give us great, simple advice.  So clearly my premise is wrong.  You can get simple, ‘good’ answers…but simple and good require context.

Increasingly, people realize that simplified, generic content isn’t worth the paper it’s printed on – or should I say, the fiber through which it’s distributed.  And technology has started to adapt.  With Web 2.0, there started to be a much longer tail of content.  With a long tail, you’re much more likely to get a match between the information you want and your specific context. And with the emerging social web, it’s getting easier to leverage an extended network of friends to get contextual answers from people that know you or at least understand elements of your context.

Needless to say, these knowledge networks are helpful when you’re deciding on what phone carrier to sign up with (T-mobile or AT&T?).  But they’re even more helpful in situations where the stakes are a lot higher.  For example, when someone in a family is stricken with a rare disease, it is inspiring to witness the degree to which caregivers use the web to find answers and support from people who have been in their shoes…who understand their context.

Bottom line: You can’t get simple and good answers unless: a) you get your answers from a trusted social network or b) you allow companies to have your personal information so that answers can be more context-based.  And while social search is great, it’s limited to a smaller subset of knowledge (unless your friends happen to be experts on everything.) Consequently, maybe it’s inevitable that people will start allowing trusted companies to be more privy to their context.

So what does the future look like?  More context will be shared.  More knowledge will come online.  You’ll get better answers.  Sounds like information utopia, right? Well, if it is, we’re still some distance away, because the technologies and interfaces that will leverage context to generate good, simple answers are more easily envisioned than built.

That said, existing technologies are beginning to offer good, simple answers – often because they are context-aware.  Contextual ads sometimes do a surprisingly good job of making advertising suggestions, and behavioral technologies can further improve on the targeting.  Netflix and Amazon have very successfully employed collaborative filters to market products to you that have been purchased by people similar to you. And Google has begun to customize search results based on your search history.

Meanwhile, new offerings being launched by a number of companies make one optimistic for the future.  The intuitive, clean decision trees developed by Hunch go a long way toward encouraging the user to provide more context about their needs. Mint made a “mint” because they did such a great job leveraging detailed purchasing history.  Geo-enabled mobile tools like Foursquare know exactly where you are before you’ve even clicked “submit.”   Shazam tags the music you are listening to the moment you’re listening to it.  And Layar similarly can augment the functionality of the eye.  It’s an exciting road ahead!

Factual hopes to be a critical tool in the generating-good-answers equation – since you can’t provide good answers without good data.  But we’re only one part of that equation, and I look forward to a new wave of companies dedicated to creating context-aware “good answer” search engines. The challenges are many, but I’m confident they can be met.

– Gil