Friday, December 17, 2010

Mental health

I was at a middle school in Bellevue today. It's striking how there are predominantly 2 types of pamphlets on the message board there: those talking about therapists/mental health and those about domestic abuse.
This, in one of the top school districts of the the nation.

Also, check this trend out. This is from all the books scanned by Google written from 1900. It clearly tells us what people are more interested in.

Expected? Or do we take ourselves more seriously than we have to?

Wednesday, December 01, 2010

Social Q&A Systems

A lot of times you need answers to long or open ended questions. e.g. "How are hiking trail distances calculated?" or "Are there any places where you can watch the 4th of July fireworks over Lake Union while avoiding the crowds? " or "Any good popular science book recommendations?" (all questions which I've had in the past).


Current search engines don't do great for these queries, since they search by keywords aim for speed. But there are a lot of Q&A systems such as Yahoo Answers, Quora etc. out there which are meant to help you with these. Having tried out a few over the past 2-3 years, here is a rundown on them:


Yahoo Answers

Pros: Lots of users, has been around for a long time.

Cons: Not really useful for genuine questions. Funny/not so useful answers tend to be voted up.

Yahoo Answers is a good example of "what doesn’t work".


Twitter/Facebook status messages

Pros: Asks questions to your friends. Decent search capability.

Cons: Asks questions to your friends, which for most people is a limited set of folks with little or no subject expertise.


Quora

Pros: Focuses of collaboration. Almost wiki-like

Cons: No user-reputation system makes answering questions less gratifying. New and hence not many users.


Aardvark

Pros: Asks questions to topic experts as determined by their internal algorithms.

Cons: Not as many users. A lot of questions go unanswered. No reputation system.


Facebook Questions

Pros: Lots of users. Uses the social graph information for question discovery.

Cons: No reputation system. Very poor search functionality.


Stack Exchange websites (e.g. StackOverflow)

Pros: Very detailed reputation/badge system which incentivizes people to give good & thorough answers to questions. Decent search functionality.

Cons: Niche set of websites for specific topics like programming, hence not generic. Not many users because of being niche websites.


So based of these experiences, here is what I think will make an ideal Q&A system:

1. Has a lot of active users

2. Has a reputation system which incentivizes good "answerers".

3. Easy to use, via Desktop/Phone/IM.

4. All the questions and answers are well-tagged and easily searchable.

5. Seeks out topic experts for answering questions.

6. Has a social element (people prefer answering questions from their friends instead of random strangers), but does not overdo it (since you might want real experts a lot of times).


Points 2, 3, 4 are achievable design goals for a startup. 1, 5 and 6 are much harder since you need a critical mass of users to achieve them, which would be tough for a startup to achieve. So it has to come from one of the big players. Or a startup bought over by a big player (Aardvark was bought by Google recently). That said, Facebook seems to be best positioned to build such a system if they wish to. Yes, it's Facebook's world, we just live in it.