More thoughts about Mahalo.com

This blog is about ambisonia.com … but I’m really enjoying thinking about mahalo.com, putting myself in the shoes of product manager for mahalo.com … and seeing how I would think/process its challenges. I think this helps in my thinking for ambisonia.com.

I’ve been very critical about mahalo.com thus far (and Jason Calacanis’ move from general blogger to pro-mahalo blogger) … but if there’s one thing I’ve learnt, it’s that websites should always be in constant evolution… and you are probably better off launching something incomplete, because you can then observe its performance to guide its future evolution. The key here is to be sufficiently ‘loose’ so that the site can become what it wants to become. I think doing this is very difficult.

Ok so Mahalo.com is a human powered search engine. Its not really like Wikipedia because the search result ‘authors’ are for the most part ‘employed’. Here’s some context:

  • Mahalo is aiming to supply results for the top 30,000 search terms. This is achievable by a not-too-big team.
  • But this fellow who created dmoz (an older human powered search) reckons that the only way human powered search engines can be successful is if their results rank high on Google… because, he states, no one is going to move away from Google.
  • To rank high on Google, you need to do your SEO (Search Engine Optimisation)
  • Jason is historically very critical of any SEO activity, he says it is like spamming.
  • Jason’s angle is: I want people to come to Mahalo _instead_ of Google (if they dont find a good Mahalo result, Mahalo will give you Google results, so its all good).
  • Others, like Dave Winer, say they cant play with Mahalo because it is too one-way… its not exciting (empowering) for developers
  • Jason says Mahalo isn’t for developers, its for the masses.
  • Jeff Jarvis says “let me recommend some links without having to create an account”

OK, so these are my feelings:

  • The top 30,000 search terms is a lot of hits. But my question is, how many times do I search a term that is in the top 30,000 terms? For me to move to Mahalo, I’d have to get at least 1 good result per day or couple of days. Say I search 20 or 30 terms every day, that means I would have to search at least 1 of the top 30,000 search terms in 20 or 30 for me to be tempted to move to Mahalo. So I say … Mahalo should not aim to supply the top 30,000 search terms, but rather it should aim to satisfy, on average, 1 out of every 20 or 30 terms searched by each individual. The big question is … would that be in the top 30,000 search terms? or the top 30,000,000 search terms? (I have no idea).
  • If Mahalo cant supply me with one good result every one or two days, I’m not going to bookmark it..
  • Jason’s expertise, his skill, is in communications shenanigans. I guess that makes him a kind of natural promoter. How is Jason’s character embodied in Mahalo? How is Mahalo an extension of Jason’s strengths. Or, more appropriately, how _could_ it be?
  • Is there room in a search engine, not for results _ABOUT_ a term, for for the most recent hype about a term (for example)? .. or the most activity around that term.
  • The page I find the most interesting on Mahalo.com is the Internet Zeitgeist page. I’ve been back there several times… to check out any new things.
  • So is it possible that when I search for ‘surround sound’, instead of getting results like Dolby’s website (what google gives me), instead of getting results that tell me ‘reference’ info (Wikipedia), but rather relevant in a ‘recent activity way’.

Perhaps Mahalo should not compete so directly with Google.

Google gives you a list of websites relating to a term
Wikipedia gives you ‘reference information’ about a term
Technorati tells you who is talking about a term
Mahalo gives you …

If Mahalo reflected Jason’s character, Mahalo would give you what is ‘hot’ right now for that term… something dynamic.

If Jason’s brain was the size of a data centre, what would he tell me when I asked him about surround sound? (he’d point me to ambisonia.com :) The Mahalo page on how to find free music on the Internet … that works because it is kind of cheeky, cutting and empowers the people who read it. Its a bit more of a ‘secrets revealed’ type thing. Buts its not a ‘search results’ style success… its more of a simple ‘howto’ success.

Ok, so the step I’m going to suggest now is… stop thinking about Mahalo as a search engine, stop casting it as a competitor to Google. So far, it doesn’t even feel like ‘human powered search’, it feels more like a personal guide. When you are lost in a city, you dont want a list of all the hotels in that city, you want a local to tell you which hotel to go to. That local wont stop there, he’ll also tell you where to eat… but perhaps more importantly than that, he might take you out that night and introduce you to his friends. oh no! there’s social netowrk in that logic.

Ok, so what is the role of social networks in human powered search? I’m actually very interested in knowing who wrote the page on downloading free music, because I’s like that person (or people) to recommend me some tunes more directly.

Combining social network with human powered search… like sending off a ‘twitter’ to your friends and they answer with heaps of good answers. Imagine if there was a system like that which could automatically record the results.

Imagine if there was a search engine that automatically log and index all the responses to someone when they twittered a phrase which starts with “where can I …”. Is that possible? Can I subscribe to all twits?

About these ads
This entry was posted in Uncategorized. Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s