Some thoughts on Mahalo
Mahalo’s out now, in alpha - the human-powered search service. It’s goal is to utilise the community to hand build search results pages with links editorially chosen for their goodness. They’re paying folks up to $15 per search results page with an option for you to waive your fee and have it redirected towards the Wikimedia Foundation.
It’s an interesting tactic although I have some strong reservations about it’s ability to be successful. If I understand how it works, it is essentially a wiki all of who’s pages rather than having information have links. If you try to search for something that doesn’t have a Mahalo page you get google results.
My first problem is this - essentially it’s only ever going to have pages for single entity searches. I.e. if you want to search for “linux” that’s great but usually when I’m searching for something I want to search for something more specific than that like “linux mbr” so all I’d get were google results. Because in the realm of what people type into search results, there are so many variations most are not going to have specific search pages.
My second problem is that it chops off the long tail in two ways. All you get are the search results some anonymous person has submitted and another anonymous person has approved. In general, I suspect, these are going to be the most popular sites on the topic - which makes sense. Unfortunately, sometimes people want more than that - I know I’ve gotten hits on this little blog 8 or 12 pages into google’s search pages. No such luck on Mahalo - the long tail vanishes from view. Beyond that the most in depth results are going to gravitate towards the most popular search results, long tail searches if they have a Mahalo page will most likely end up being limited in scope and vision as they are worked on by fewer people.
The third problem is that approved search results for relatively obscure searches may be refreshed very slowly. So, if I’m an expert in some wierd subject and I go in and do some great pages for a few terms on that subject and then get tired of doing it those pages will continue to give old results thus actively hiding from the public any new great sites on the subject.
The fourth problem is that search results are a very, very hotly contested area. Companies rise and fall with their placement in search returns. I question the ability to keep a huge repertoire of links ethically clean. Who can vouch that the person who submitted a good search page didn’t take payment to include a few iffy ones?
Those are some of my main concerns with a product like Mahalo. Obviously it’s very early stage right now and they may have already addressed or be in the process of figuring it out. To me, though, at this point it seems like a more buzz friendly version of About.com or perhaps a more linkified version of Wikipedia. In my mind, at least, this is more of a competitor to those sites than to a search engine like google, that is more likely to be a search result than a search engine in the way I would use it. In many ways, it’d be more useful to me as an add on widget to Wikipedia than a standalone site.
I’m curious to see how it progresses and if it make it.
