Jason Calacanis Launches Mahalo Today: Human Powered Search
Michael Arrington
127 comments »
Jason Calacanis, formerly of Weblogs, Inc. and more recently an Entrepreneur in Action at Sequoia Capital, will launch his newest startup, Mahalo, this afternoon at 3 PS PST. The site is password protected until then.
Mahalo is a search engine, and will join Powerset as the more interesting new engines to launch in 2007.
The service has features that are similar to the new Netscape news finder product that Calacanis launched last year at AOL: expert guides will determine the most relevant results. The main search results are provided by guides (Mahalo employees), who find relevant results for search terms. User submitted results are also included.
The primary results for search terms are included at the top, in a “top seven” area. These are hand picked results from the guids that should all be good results for the query (see screen shot below of results for “Paris Hotels” - click for larger view). To the right of the results are “Guide Notes” which include additional information including relevant additional searches and “Fast Facts.” In the case of Paris Hotels, the “Fast Facts” include the country, language, currenty and telephone country code.
Additional hand picked results appear below the Top Seven, and Google results round out the query.
Also in the right sidebar is a place for users who’ve registered with the site to submit links relevant to the query. The more users who’ve submitted a unique link, the higher it appears on the list. Mahalo will be looking for fraud in this area - if a guide determines a link to be spammy, they will ban the link, the user and the user’s IP address “forever.” However, if a link gets enough votes and is determined to be relevant by the guides, it will move over into the main search results area.
Each search page also has a discussion/forum area, where any registered user can add their thoughts to what’s included on the results page.
Mahalo has 40 full time guides today and have created 4,000 results pages - each of which will serve approximately 12 various queries. Calacanis says that the guides are steadily improving results and adding more queries - they expect to have 10,000 by end of year, and 25,000 by end of 2008.
They are really focusing on top search terms, which they obtain from a number of search engines and other sources, Calacanis says. If a user makes a query that has no Mahalo results, they are given Google results instead. Calacanis stressed that it is going to take them a couple of years to get really deep results for most queries - until the end of 2007 the site will be in Alpha, and will be in beta for all of 2008. “It takes a few years to build a really compelling search engine,” he said.
Mahalo has now completed two rounds of financing. The first was led by Sequoia Capital. The second was led by Elon Musk and News Corp. The size of the rounds have not been disclosed, but Calacanis says he has enough capital to run the business for up to four years without revenue.
Disclosure: We have no financial interest in Mahalo; however, Jason Calacanis is our partner on our upcoming TechCrunch20 Conference.
I tried to access the site to try out one of the available 12 queries, but it is password protected. So i headed to Google, who offers a googleplex of queries.
yes, it will be live at 3 pm…see above.
Sounds more like a directory than a search engine
4,000 results pages they created so far, I guess I can switch from Google if my query matches one of those 4,000. Idea doesn’t make sense, why would I switch from Google to this? Are these guides experts or are they just doing Google searches and picking out which results they like and then adding it?
It looks… cute and pretty.
From what I read, the concept seems to be a hybrid between Ask, About.com and Answers… Of course, I haven’t used it yet. But Mahalo?
The idea is that you are going to get Google results if you are not asking for something they’ve created results for. So if you like their results for the queries they are adding, this might be the search engine for you.
That is a great question how are the experts finding the content?
What they should add is some community feedback to round out the “expert’s wisdom”. I mean, it’s interesting if someone believes that a search is relevant. It’s far more interesting if everyone believes it is.
Again, I am always in the Swicki camp on this one. The most effective way to use a search engine is to have users decide what is important to them. The wisdom of the crowds might be flawed, but it’s not difficult to tell if the information on the page is actually representative of what you were searching for.
It’ll be fun to see how good these experts are, but I’m hoping a future version of the product expands on the idea a bit.
How are they getting the results though? It doesn’t sound like they are experts but just some average Joes doing searches and just adding whatever they feel fits. That’s not a very good selling point because I could get the same results from doing a query with Google. I do agree that it sounds more like a directory then a search engine. What’s next, people are going to pay to get in those top results like yahoos directory?
Just a guess, but it sounds like something the navigators from Netscape would be doing.
Steve - there is a discussion area for every result page…as mentioned above.
Yes, there is a discussion area. I guess what I was seeing was more akin to Digg than to that. Where users opinions actually figure into how the results are sorted.
so a ChaCha? Squidoo’s Lens model wasn’t very popular either. Regardless, the point is to get people to use your search engine so that you can bank off of the AdSense revenue. Whatever tricks you can do to get people to do that is all that matters. It used to be sites would give out prizes for you to use their version of Google.
so he recreated the yahoo directory.
do you know why the yahoo directory was killed jason? too much f#%$king content for any team of editors to handle. oh, and dead links, they tend to be a problem after about two weeks as well. then you have to rebalance your ontology every six months.
PAIN IN THE ASS
algorithmic search is JUST BETTER at solving this problem
looks like another me too flop
I think the real plan is to SEO the hell out of this hand crafted result pages so that it appear for their respective queries in google… Jason is smart enough to know that it will never become a type-in destination site on its own.
That’s not a bad plan at all. Jason probably has figured out that this isn’t by any means a Google killer. It is, however, another option.
If it is positioned as a meta-result; for example, you type in “paris hotels” and get a hand picked list of the “best sites” on paris hotels, and you can get this to show up near the top of the SERP — you certainly have something interesting.
Regarding “they are given Google results instead” - is there a formal partnership agreement with Google there? Given the involvement of Sequoia Capital, aka. Google buy-out Sugar Daddy, I wouldn’t be surprised if there is indeed a co-operative arrangement.
This is the “FTN” Friends TechCrunch Network. Once you are a member, your posts will never have negativity (see guy’s truemors), you will get immediate first slot status and will always have exclusive here on TC.
Naturally Cal is in the FTN because of TC20. I wonder if we will see bashing of any of the companies that participate in the tc20 like when yahoo helped crunchgear on their home page.
Such an interesting FTN network there is in silicon valley. It’s too bad we will never get Mr. Arrington’s real views on this or any FTN member site. That’s what made TC great.
As the web continues it’s explosive growth I think people will search out services from companies like Mahalo, that provide the filtered information they’re looking for. The content for these type of sites is limited compared to google, but the content you find is relevant, and if done right, unable to be gamed.
Looks like it’s built to flip, and will be wholly dependent on Google results to bring in traffic as its too small (and unscalable) to bring in enough traffic on its own. If Google does not rate their pages highly, they will be DOA.
That seems a bit unfair. The article didn’t read like a rave review of Calcanis’ new prize pony. Now, as far as Truemors is concerned — that’s a slightly more interesting story.
I think the idea needs work, but this is before the product even launches. My main problem with it is that every time a company relies on “experts” to decide on relevance, it is always slow and always lacks the ability to adapt.
I think that’s what makes Web 2.0 such an interesting movement. Even if the people are almost always wrong individual, on average they make pretty decent decisions.
Until, of course, the SEO experts start gaming the system. Hmm…
#22 Steve - “seo is dead” and “seo is shit” from calacanis at ses chicago and on his own blog numerous times. So I assume SEO will play NO role in this program.
What I meant was that if the system was more like Swicki, then it leaves it open to external marketers gaming themselves to the top.
Good point though. SEO probably won’t figure into this.
Now I’m even more interested in what the plan really is.
Why would you assume that a few attention-baiting blog posts would interfere with BUSINESS?
Really great idea!
Can’t wait to see it (for some reason the password is still on).
If anyone can pull this off, it will be Jason…
It’ll definitely be interesting to see.
How is this a bad move?
40 x 20k = round to 1 mil /yr outsourced grunts doing directory work.
slow burn and max publicity ….
Hasn’t the “human powered search” been done before? ChaCha claims to be human-powered, as well.
It will be interesting to see how this all plays out and if Mahalo will be able to handle the traffic.
I just did a search for “search engine optimization”…funny how Thomas the Tank ENGINE is related to search engine optimization. My son loves Thomas but this is kind of ridiculous.
Looks like a slightly better Squidoo.
Its open, and its pretty slick!
Going take some time to mature the results but I’m sure Jason has hired the right people for this job. Very nice that can recommend a link, but this is only an option for existing pages. Hopefully they will look through the logs and see which missing pages pop up the most.
Interesting. I’m a big fan of the development of search technologies. I think we are only just beginning with possibilities and that there will be more to come in this space.
As for the site, it looks well thought out.
Ha!
I can’t even find what I’m looking for. But it gives me Google results.
Fantastic. Hey, I have an idea. Why I don’t I just use Google instead and save myself the trouble of looking at the crappy logo with the ‘we-think-we’re-so-clever’ alpha attached? Sheesh. Calacanis, will you ever learn anything about branding or style?
No digg.
Interesting, it looks like they are only looking to do the top 10,000 search terms. This might be very doable. Though, there is a question of whether the service will change with the content.
Yeah uh, I don’t think I could find answers to “Parser Error Message: Can not access file ‘System.Web’ because it is being used by another process.” or any other random problem I might have like I could with Google.
This is probably more for entertainment than actual use.
I think I heard recently that there are 1 billion searches per day on Google and over 25% are new keywords.
How does “just the top 10k searches” fit in with longtail thinking?
Wait a Second
“They will ban the link, the user and the user’s IP address ‘forever’………”.
Is it only me who sees a flaw in this???????What if someone working in IBM corporate office in NYC submits a spammy link.
Will all the employess , who share that IP address within the network, get banned?????????
Any thoughts??????????
Will About.com/NYT be after them for the use of the term “Guide” as someone who collects and organize information on the Web?
Unfortunately Mike, I think you missed a few critical points which I included in my review on Cn:
http://www.centernetworks.com/.....cal-review
1. They state they are only creating pages for the top 10k searches. What about the rest? They get the same ole Google. Of course Jason did a smart think and listed all the silicon valley peeps. I can only imagine that CK Sample (whomever he is, I think a buddy of Jason’s from AOL?) is not in the top 10k searches. So exactly how does one get a page made?
2. This is the same as AboutUs except prettier. And on AboutUs, I can edit my page. So I guess the score is 1-1.
3. This is not a human search - its more of a directory. For example, using the current definition of search, if I put “techcrunch” into the search box, I should see results that are based on something for the term across the live web. Instead I get a page about Heather Harde and Michael Arrington’s blog. That’s why I call it a directory not a search engine.
4. How will they keep this maintained? What’s current today is out of date tomm. Goes back to #3.
5. I might be looney, but is this not the best SEO play out there? I go into more detail on my post on Cn, but this affords a chance to get major play in Google for every term they write about. I said the same thing about AboutUs (ha) last year. So while Jason has said over and over that SEO is dead, his product is at least partially a SEO product. Interesting.
Steve S,
Your first post about the importance of SERPs was probably correct. I think Jason has seen the success Wikipedia has had and wants to duplicate that with Mahalo. And since he is paying contributors he doesn’t have the problem of being unable to monetize through ads that Wikipedia does.
Can someone with the TECHNICAL KNOW HOW resolve the following paradox from Mike’s review of this new Hawaiian baby
“They will ban the link, the user and the user’s IP address ‘forever’………”.This is directly from Mike’s review
What if someone working in IBM corporate office in NYC submits a spammy link.
Will all the employess , who share that IP address within the network, get banned?????????
@42 Yea, it sounds as close to being right as I’ve come up with so far.
It looks like it can be useful, if you happen to be “surfing for fun”. The top 10,000 searches are the top 10,000 for a reason. The majority of people casually using the system will amble over to those searches at one point or another.
The problem is that it doesn’t seem like this system is designed for casual use. It’s a specialty search service in the same way that imdb is. Think of it as akin to a buzz tracker.
It should be an entertaining diversion though as it stands.
The name does not sound to me as a better brand, so would suggest that he buys and name the site http://www.mediarati.com
no results for “porn”, i could get notified if results are in future available, but i need it NOW …. jason goes deadpool on this one.
Steve S,
I don’t think Jason believes that Mahalo will ever be a destination search engine. Rather, like Wikipedia, IMDB, TV.com and the like, Mahalo pages will primarily be found through Google, Yahoo and Microsoft SERPs. The fact they are concentrating on the top search terms supports this.
Looks like Jason should return to being an “Entrepreneur in Action”
this is frankly, a dog! you might as well throw this one in the deadpool now.
Perhaps Jason is too young to remember but Yahoo used to be a human edited directory… in the last century. google built a better mousetrap.
“history is doomed to repeat itself”
@47 Yea, I tend to agree.
And what makes these people experts because the Navigators on Netscape certainly aren’t what I would call experts at all. The results are so locked down, its like you have to be a friend of the experts to get listed. I mean really look at the Weblogs Inc sites listed prominently on relevant topics. What makes their coverage any better than any other sites out there?
This is lame Jason
I hope it stays fair and non-biased just like ODP.
/sarcasm
I’ll say this for the setup - at least the fallback Google results know that when I search on ‘art’ I don’t want Artichokes or Articles or Ron Artest, but things containing a-r-t
Sheesh, after all these years, artichokes. It’s like 1994 all over again.
#50 tony - that’s the point, or did you miss that part
It’s nice to see Calacanis has the connections to get something like this funded because this idea isn’t really groundbreaking for the most part. I’m interested in seeing how his model based on human editors is going to scale. I agree with an above poster that this thing is being built to flip.
SEO has definitely got to be one of their top strategy for traffic.
Of course Jason’s defense will be he is against THAT certain kind of SEO and not SEO that is “naturally” good(*cough* mahalo). I’m not saying I buy that but okay.
-Zaid
The way I see it, there’s room for Mahalo.com because it’s not just a search engine but it is a directory. A directory of everything popular, at least for now. I think it will be a major chore to keep the links active and the pages up to date but thats why he has 40 guides and I’m sure he’s still hiring more. The problem that this solves is relevant content that is getting more difficult to find on Google, Yahoo, Ask or any of the others. Who wants or needs 500k search results when the top 7 is probably all you need or at least it’s a great start. I like the concept and I could see some real vertical applications for this type of search directory.
yeah more of a directory
- 40 Full time employees / that sit there and make ”fav lists”
- the idea is broken ……
- Can executing a broken idea perfect / make it work?
- we’ll see
Sounds a lot like Bessed.
I think the best thing that will come out of this is a really interesting news bite and a pretty cool directory.
Scale is definitely not the main problem, he’s only after the top 10,000.
The problem will be keeping things relevant. If the major links are Wikipedia and Imdb like in the case of Star Wars, then OK. For more obscure topics, it’ll be more of a problem.
oh no, don’t tell me..what the hell
translation: this sucks.
Wow you people are negative. It looks pretty helpful to me. Instead of sorting through all the crap, they have what I would likely be looking for already put together. If you don’t like it, don’t use it.
After looking at results for a few items, while it may have more relevant results than Yahoo or Google it’s not as information rich as Wikipedia. It will be hard for a staff of humans to keep up with all the humans on the web.
hey i have one question only…
WHAT DOES HAMMER THINK OF THIS?
lol
But why do I care what these people think? What makes them experts? What makes their opinions matter at all? What makes Mahalo matter?
Hey mahalo, your logo needs better anti aliasing, looks pretty jagged, and it is your first impression…other than that I will give it a try.
Mahalo — the Hawaiian word for “monetize”
Calacanis has been tipping his hand about this on his blog for months, every time he tweaked Jimmy Wales for being squeamish about monetizing Wikipedia. This is his attempt to steal that prime Google front page ranking from Wikipedia and the as-yet vaporware human search engine that Wales has been touting forever. Good on him.
How do they ensure that the results stay up-to date? With the Paris Hotels example, what happens in a year when the hotels they initially selected have different star ratings, better sites have appeared, or the prices have all tripled due to increased popularity driven by Mahalo?
http://www.leveragingideas.com/
dmoz.org meets web2.0
Good to see an up to date directory with funding behind it though.
If you guy… Reminder never put flower logo on your startup. People will think you are a f*gg*t.
Is Jason for real with this?
http://www.mahalo.com/Arctic_Monkeys
This is just a link farm….there is no REAL information here. Why would anyone use this over Wikipedia or just Google?
Here is the recipe to this site:
1 part Digg
1 part Google
1 part Wikipedia
Add a little unoriginality and you have “Thank You” in Hawaiian??
No wonder they gave this site such a terrible name, they better be thanking their investors for the money.
It just goes to show that it’s about who you know, not what you can do.
how annoying is that person from center networking? Hey Stern, why don’t you write some original stuff on your blog?
what a disappointment! Sorry.
Typed in “yahoo” and get
MediaWiki 1.9.3
Please set up the wiki first.
No idea what the fk is going on.
Back to the drawing board.
I should pay closer attention (or read the whole article). I thought for some reason it was natural language search.
I don’t know - I think it has potential but it’ll depend, like any of the other zillion start ups that have been launched. I would ditch the flower, also, and tag line.
http://www.mahalo.com/Special:.....;go=Search
After taking a look, I think its a concept that will work better with http://www.mediarati.com cos what I see here is more of a reference to sites, blogs and forum where one can get the relevant infomation. and this makes more sense with mediarati. Maybe I rather start looking for funding.
Thanks Jason for the first mover intiative though.
I usually refrain from mentioning my own company in TC comments, but in this case, I have to say ‘been there, done that’ - literally. Wink 1.0 was people powered search (Wink changed to people search recently and it’s taking off much faster).
We found that with thousands of users and well over 5 million user-selected links we were just getting started. Results were pretty good for a lot of queries, but the tail is *really long*. As anyone with access to a query log knows, there are a small number of common queries followed by millions of queries that all seem to be different. 10K hand-crafted queries is less than 1% of the coverage of Wikipedia - and that is content, not search.
Not sure that Calacanis has any actual search technology - this is more of a directory or guide - like About, BlueDot, Squidoo, or Fanpop. Could SEO well if they can get these guides into top results - but can be hard to SEO pages of links. Wish them luck, though - good thing they’ve got 4 years of cash.
Why does more results linked to google search mode?
People see it. A spam site. I think Jason Calacanis need to work harder on that search engine. I wonder if they use PHP.
if you have VC money.
Never… Never… Never use PHP script to write on search engine It’s not professional. Always use ASP, ASPx, Perl, and other scripts to make search engine.
The real question, is what will it take for people to break their Google habit? I remember my first Google query almost as much as the first time I…I…I mean, seriously, though. Give me something that will do to Google, what Google did to Altavista and then I will be impressed.
Sorry, your search engine couldn’t pick up the correct stock quote.
Not Human powered search.
Sooo busted!! :p
Human powered? How about alien powered. I’ve being seing this tags everywhere on the Net. Have they think about trademarking the tags “Human Powered”. I wonder why they didn’t until now.
Look like “Intel Inside” is the only tags alone being trademarking.
Thank u ,so interesting it is
it probably wont work… with semweb around the corner this is NOT a long term vision.
I didn’t think Google listed other search engines results to start with? If I were them, I wouldn’t count on Google traffic for too long, unless they decide to call it what it is, a directory.
Humans makes more mistakes than machines.
good lord, why are you blowing sunshine up this company’s ass? honestly, are “internet users everywhere” really searching for “michael arrington”?
I like how at the bottom of the site they say “What is Mahalo? Mahalo is the world’s first human-powered search engine.”
Ummmm… didn’t ChaCha do it first??
I temp’t at Looksmart when the bubble unemployed alot of us in San Francisco.
Hand curated search. This feels like Looksmart with better 2.0 tools.
Music results (at least at the band level) feel like FoxyTunes Planet.
Looks like a scraper to me. Might end up in the Google SERPs just like the junk they supposedly want to filter out.
Reading through the information about mahalo.com it somehow provokes me to mention about http://NosyJoe.com, which is yet another –not as ambitious but– interesting example (although still in private beta) of a social search engine that relies on humans (nosy preferably :-)) to build up its index massive. It offsets the low quality content by extracting only the meaningful and grammatical components of the content and further organizes the information into phrasal and keyword based tags. From what I have seen on the site so far it also appears to me that it mashes-up, ranks and shares the information in an interesting manner, although I do not quite understand how, by applying a set of algorithms and user patterns, as they claim. The Nosy Joes also seem to believe in the fundament that humans do it better
While I understand wh