Posts Tagged ‘intent’
24th January 11
We recently posted a rant about Quora that generated a lot of conversation. One of the comments was by Leslie Barry, the founder of Iphso. Leslie made the intriguing argument that Quora actually gets closer to question-and-answer nirvana than any other service: identifying intent. Here’s his explanation of what we’re just not getting. We couldn’t be happier to hear his perspective and would like to thank him for generously agreeing to guest post.
*** *** ***
Author: Leslie Barry (@LeslieCBarry), Founder of Iphso
What is intent?
According to The Search by John Battelle, the holy grail of search is to interpret the user’s intent and direct them seamlessly to the content, or ideally, provide the answer directly.
If we assume that intent is not WHAT I’m asking, but WHY I’m asking it, then I believe that Quora is closer to solving the intent of search.
So how does Quora get us closer to it?
Quora’s approach is to get the best qualified people (through credentials or experience) to create some rules (boring to some, but necessary to prevent chaos – even Wikileaks has rules), then leverages the serendipity effect to overcome the constraints of similar services, like LinkedIn, to refine the best questions to elicit the best answer.
What makes Quora unique is that it uses serendipity more effectively than other services.
What is the ‘serendipity effect’?
Wikipedia defines serendipity as ‘a propensity for making fortunate discoveries while looking for something unrelated.’
An example is the real-world benefit I’ve experienced from social media, specifically Twitter, where I meet useful people that I would never know existed without a serendipitous network – i.e., people that I didn’t know I needed to know.
Quora is leveraging this extremely well: connecting the right ‘people I didn’t know I needed to know’ to clarify intent of the question, and as a result, shortening the path to the answer.
Intent is only relevant past a certain threshold of question. Something as simple as ‘What is the capital of Maine?’ is clearly better answered via Google. No rocket-science behind the intent there, but asking something slightly more complex and ambiguous like ‘How do you consume news? Has this changed in recent years?’ presents a greater challenge. When you Google it, Google assumes:
- The question is correct
- Keyword/location/context matching is adequate
So here is what Google thinks the answer is:
…and then Bing:
They both completely missed the point without the Quora results!
Conversely, this is what Quora’s users think the answer is:
Clearly, Quora is more efficient at interpreting intent.
This is because Quora doesn’t assume the question is correct. Instead it provides the ability to ask a question and have it clarified and modified wiki-style to help shape and tease out my intent. Often as a user, I’m clear on my intent, but am not the subject matter expert and therefore unclear on how to frame the question. The iterative, near-time editing of the question helps solve this issue. Also, Quora doesn’t focus on keyword/location/context matching like a mind-less search engine. Quora’s process of refining the question by subject matter experts eliminates the extra steps of sifting through multiple, non contextual answers.
And yes, there are many planted questions (Google Link-bait pages, anyone?, LinkedIn self-promotion?, mindless waffle on Yahoo Answers?), and self-promotion, but so what?
All I care about is the quicker path from question to most valuable answer that addresses ‘what I meant’, not necessarily what I asked. My intent.
Why can’t Google or Bing use other people’s answers to opinion-type questions to decide the most relevant options to serve you? I think this is because of their strong focus on the search algorithm, which values content over context and popularity over intent. Also, Google is one-way traffic, without an iterative, refining feedback loop; we have to stumble blindly along hoping for one or two interesting search results out of ten or twenty.
As a result, it’s Quora, not a traditionally defined search engine, that’s helping us take a step towards the holy grail of intent.
It’s not perfect, but it’s challenging our approaches and thinking about teasing out intent from users. Maybe Google indexes them and learns from them? Once again, it doesn’t matter – we will have taken a huge leap from accepting that indexing and search is better than curated, considered, intelligent answers.