It is getting harder and harder to find things online these days. Most people default to Google and Google returns a lot of results. And we have to sift through them. Yes, definitely a lot of good results are on top, but a lot are not.
Recent entry into the search space by Mahalo is a perfect example of why there is an opportunity. Say we search for a movie Bourne Ultimatum. Here is what Google comes back with:
And here is what Mahalo gives (btw, tip to Mahalo team, learn how to use div element and make pages resizable, its difficult to take a good screenshot):
Mahalo’s results are better for two reasons: they are hand-picked and they are better formatted. With Mahalo a person does not need to sift through links, most likely the best link is right there.
The secret here is semantics. Mahalo workers are people and people know movies and they know good sites about movies. And this is what lets them pick good links - understanding that Bourne Ultimatum is a movie. Google also “knows” that it is a movie, but it certainly does not take full advantage of this fact.
But Mahalo has another problem, perhaps its forever fatal limitation is freshness of the content. Because it is done manually by people Mahalo has no chance to stay on top of rapidly growing web and even blogosphere.
And the problem is not as acute for web sites, it is much bigger for blogs, which Mahalo is not even addressing yet. How do we find the best blogs about this movie? The Google Blogsearch does this automatically, but just like Google returns a list results that are not as filtered as Mahalo.
We think we developed nice, and straightforward solution to this problem. For each category in BlueOrganizer and SmartLinks, for example movies, we researched a list of best blogs focused on movies, primarily reviews. We then used Google Custom Search technology to create a specific, vertical search engine for movie blogs. And then we exposed this search, simply as an action in the SmartMenu and as a shortcut in the SmartLinks pane.
Results? You will soon be able to see for yourself, but our tests show that the number of results that come back is small and most of them are of very high quality. The combination of contextual understanding and targeted search is very powerful. This is what Mahalo team is doing manually is basically that.
Of course our engines will present the results in the same format as Google, so Mahalo will still have that one. But in terms of quality of the blogosphere search, these vertical searches are going to be pretty great. More examples are coming soon.
{ 4 comments… read them below or add one }
Re: the comparison of Mahalo and Google: Even though Mahalo is cleaner, I have to say that I think I prefer Google’s results, and that’s all about perception. A few reasons why…
- First, Mahalo really isn’t giving me anything I don’t already know about. I already have a personal association and a comfort level with Rotten Tomatoes and Wikipedia, for example, so I don’t know why I wouldn’t go directly to those sites to find info about Bourne. Mahalo seems like a middleman in this instance.
- Second, Trust. Addressing your comments, yes, manual updating is a forever fatal limitation. It would be hard for users to build trust with a search engine that is manually generating results. It’s not scalable and they would have to have some seriously consistent positive experiences with it that are above and beyond the results they’d find anywhere else, which in this model seems unlikely.
In addition, Who’s actually picking this content? When it comes to filtering content, non-transparency is an easy way for users to disconnect because they don’t trust the filter. What am I missing? What is out there that I can find that you (Mahalo) wouldn’t. It feels incomplete… like a wikipedia with only links and no community.
As a sidenote, if you want to see a good manually-run content aggregation site, check out http://buzzfeed.com/ …the key to this is that they’re offering something of value above what other sites are offering: an editorial voice and consistently-satisfying content. This builds a user trust with the filter(ers).
- Third: Google is serving up additional information in an instantly gratifying way. I can already see the movie’s length, 2 different ratings systems, its categories, a review tally, and can scan the dates and popularity of links, a photo for related information, and brief bits of information on web sites that are out there. It’s not as actionable, perhaps, but it feels like I’m getting more.
In comparison to that, Mahalo feels like it’s actually blocking me from information, even though it’s provided me a cleaner view of links that could actually be more useful. (That, and design-wise, the text I should be clicking on is smaller than other text on the page.)
Instant gratification and perception are the keys in improving Mahalo’s displayed results… those aspects are something we’ll definitely have our focus on at AdaptiveBlue, as well.
Hi Rion,
You are absolutely right from the perspective of what I call very sophisticated user. You love to have control and choice. A lot of more mainstream audience may not need this. Or may not need this right away.
Mahalo is sort of like Wikipedia, a good first stop. And then Google to get more details. The thing is that a lot of times, first stop is all we need.
Re: who picks the information. Yes, definitely there is a trust issue. But this is exactly the same with New York Times and CNN.
While I do agree that Mahalo is certainly prettier than Google (what isn’t?), I feel like it’s holding my hand a bit too much. Perhaps the example being used is a bit too general. Would most users really only type “The Bourne Ultimatum” if they wanted a movie review or a DVD? I think even middle-end users now have go-to sites the same way they have favorite brick-and-mortar stores, so in a sense it is indeed a middle man. And the low-end users who can really benefit from Mahalo’s semantic search results? These are the same folks who think Yahoo! or Google *is* the internet and know no others. But I suppose this is Mahalo’s marketing challenge.
And yes perhaps Mahalo will excel over Google at packaging and presenting general search queries, but what about when one searches for “deadwood season 2 episode guide”? Isn’t this an example of more common usage of search engines? In this case, Mahalo defaults to…Google search results.
To each his own, I suppose, but 99% of the time Google is finding what I am looking for. Perhaps Google should just buy Mahalo and integrate its functionality. It’s certainly easier to spell.
Hi Richard,
I agree with you in terms of limitations. My point was that sometimes quick and filtered results are good. We already have Google if you need broad and comprehensive. So for some things it is nice to have quick and simple and filtered by humans.
I do agree that scaling this and staying on top of the web is challenging if not impossible, but it is probably not the goal. Lets wait and see what Jason and his team have in store.
Alex