(shot by Lenslinger)
Man, if Jason McCabe Calacanis is nothing else, he sure as hell is one smooth-talking dude.
Ever since Silicon Alley Reporter shut down due to the crash of the web agency industry he’s been stitching together the work of other people into something ripe for purchase. Funny how a business (SAR) that pimped those very same web agencies, driving up their visibility & rates and SAR’s advertising dollars, died the same death.
I heard a lot about Calacanis’ latest project, Mahalo, over the past 6 months, primarily through friends conversations out on the left coast and the occasional Scoble / Techcrunch / Winer ruminations. But when listening to Calacanis present Mahalo to Ed Cone and the audience this past weekend at ConvergeSouth, I have to admit I was a bit taken aback by some of the claims in his elevator pitch.
When contextualizing the features of Mahalo in the search world, Calacanis spoke of Google’s search algorithm (PageRank) in the past tense. He continued on, bloviating about the old way of presenting search results, where a machine based relevancy off of explicit criteria found within the markup, such as keyword matches with domain names and titles and the number of links to individual domains. This, Calacanis argued, creates relevancy that cannot be trusted as valuable, because SEO is a game and results are flattened out without any consideration of quality.
Mahalo is apparently different because Calacanis has hired a team of employees experts to create result pages that enhance the results of the zeitgeist of society (through the top 25,000 search queries). Forget the subjective argument surrounding “expertise” and answer this one question for me:
How does this make Mahalo a search engine?
The way I see it, Calacanis has created a super-collaborative link blog. Sure, there are community elements to it and there is a search box up top, but authors (or “guides”) are simply culling together their top link choices to give further context to an idea, issue, topic, place, etc. Mahalo isn’t a destination search engine; Mahalo pages are built to show up high in… you guessed it, Google search results.
None of this is bad, but it doesn’t make Mahalo a search engine. Or at least one attempting to compete with Google.
See, the vibe of the pitch keynote leaned pretty heavily in the direction of Mahalo being the future and Google being the past. Being that Mahalo is human in a miscellaneous fashion, I won’t compare it to Yahoo! back in the day — a site that attempted to categorize every known site into a master ontology. This is much smarter, as the well structured page titles, super-relevant links and structured data makes Mahalo pages ultra-ripe for a Google crawl and a well placed search result.
Calling a spade a spade — an SEO optimized link blog, trying to gain top spots in Google to cash in eventually on AdSense or similar — isn’t what CEOs do in this world.
To front like Mahalo is revolutionary in redefining search or that it will become a destination search engine — particularly, one that can marginalize Google’s PageRank algorithm through human expertise (btw, this very same algorithm is the one exposing these secondary result pages to the general public in the first place) — is a bit of crazy talk.
For shits and giggles, say that Mahalo succeeds in creating result pages for the top 25,000 search queries from Google and Yahoo!. Then what? How many “expert” monkeys are needed to not only scale to meet the demand of the ever shifting zeitgeist, but to maintain pre-existing hand-linked search result pages?
How long can Calacanis’ pet example, “Paris hotel,” stay relevant as the months and years creep on by?
Or is that the phase in Mahalo’s strategic plan where the real crowdsourcing kicks in?
Interesting stuff to ruminate over, but to lead off a conference hyped as “Creativity online for all people”…?
Jason isn’t charming enough to get people to whitewash his fence out of love. Nothing can change that other than a CEO-ectomy.
I agree with your perspective that it is a SEO link farm with a smattering of original content. Ironically, mostly it is a bunch of links researchers have taken from doing a google search or hitting up wikipedia.
In the good old black hat days manual and automated variations of this worked to get traffic through organic listings from google…now that doesn’t work unless you have the spin machine going in overdrive and building “trust rank”. Calacanis by being so audacious is generating a buzz for this product which will build his “trust” (algorithmically speaking)
The irony is for all the spiel about google bashing Mahalo is pretty easy to game.
yeah, calacanis knows what he’s doing. and quite honestly, it doesn’t really bother me. it’s just very transparent and i’m amazed how he can keep a straight face through it all. i mean, dude is dubbing his concept of “web 3.0″ as “the expert web“… but as jason will tell you, he’s from brooklyn.
(smacks forehead!)
how would you game mahalo?
How would you game Mahalo? People are easier to game than an algorithm. People can’t be reprogrammed at a moments notice to make them more resilient to bullshit
interesting angle… i never thought of wining and dining a wiki nerd to allow a page to stand, so the thought never even crossed my mind that these “expert guides” could be bought. good point.
Mahalo for the feedback. Couple of points…
1. SAR didn’t shut down, we rebranded it VentureReporter and sold it to Dow Jones where it is still, amazingly, still operating. We worked really hard to save the company as it went from 70 to a dozen or some folks and I’m very proud of that effort (although I probably should have went to Thailand for a year like everyone else!).
2. re Mahalo being a search engine or a directory I wouldn’t worry about that too much since it’s a little of both. We present result in a search fashion, but they are organized as a directory. Folks can surf the directory and discover stuff, or they type a word into the box. So, if you want to be technical we don’t crawl the web with machines, but rather humans, so if that is the definition–machine crawl–then we’re not a search engine. If you look at it from the users perspective many don’t realize we’re human-powered at this point so they are experiencing search when they use the product.
3. We’re also using social pieces to build these pages in the form of user submitted links. So, that’s phase two of Mahalo… you’ll see more expert+social interaction in the product in the coming months.
4. I didn’t do a demo of Mahalo at the event and I tried to keep the conversation to the big picture so as to avoid the “it’s spam!” label being put on the talk. Obviously that didn’t happen for you. Frankly, feels like for a couple of folks if what I’m working on even comes up in a talk I’m spamming folks. I find this odd since a) i’ve been speaking at conferences for 10+ years and never had this issue except the Winer-ing and b) most folks come to conference to hear about what people are working on. I get invited to speak at conference because folks want to hear what I’m doing, and I myself go to conference to hear about what people are working on. If I don’t like what their working on or I’m not interested I do email… no harm, no foul.
5. You won’t be able to game Mahalo to any large extent–certainly not like a machine search engine. Our search results are looked at by three different people on our payroll at a minimum. Any link that is included has to be at least “very good” to get in the index. Also, the amount any spammer would pay to get in the index is going to be less than the value of the relationship with Mahalo of the person they are trying to bribe. In other words, the $100 to $250 that someone might pay to get their link into Mahalo is not that much for one of our people, and our people are good people hired because they believe in the mission. They also have in their contracts that they can’t take any money… so, you’ve got as much chance of bribing one of our people as you do a reporter at the NYTIMES. Could it happen? perhaps, but we would find out about fairly quickly and resolve it. Also, it would impact a couple of links out of millions on the site. We use wiki software so every single edit is tracked. If someone does something bad we find out instantly and check all their other edits (like Wikipedia does). Also, we have a monetary relationship with everyone contributing (unlike wikipedia). We know their names, where they live, their social security numbers… it’s a fairly trusted relationship.
[ got cut off… posting end again ]
6. In terms of the value of Google/Yahoo/Ask vs. Mahalo look at any of our travel, auto, product, health, etc. pages side by side. We win every single time–hands down. This makes sense because we start working when Google stops. We START with a Google or Yahoo result and spend 5-10 hours making the result better… it has to better by default. How much better? Well, we’re less than six months old and our pages are the best in the search business hands down. The only issue is how many can we build, how fast and how updated can we keep them.
7. In terms of how big can it get, thanks to the Greenhouse (where the public produces search results) we are moving faster than we ever thought we could. We thought we would hit 25,000 at the end of 2008. We are 18k pages right now! We are going to hit 25k by the Q1 of 2008.
Also, our team is not a bunch of “monkeys,” they are hard working, real people who are trying to make a living by help people find quality information on the web. These are good people doing good work and I would ask that you have some respect for them and their effort. We’ve got students, work from home parents, and normal folk trying to make a little extra money to take care of their kids a little bit better building Mahalo search result pages. Try to have a little decency.
If you want to attack me for being a relentless promoter and rabid entrepreneur fine, but the folks who work for Mahalo deserve all of our respect.
It wasn’t long ago I was defending bloggers who were being called monkeys by idiots who were threatened and confused by their good work. You’re a blogger right? How would you like being called a monkey for your work?
best, j
thx for the thorough response, jason. though i’d appreciate it if you didn’t make me throw-up in my mouth with that greeting of yours.
i didn’t call any of your “experts” monkeys. to your point:
i implied that in order to scale to your vision that the team would in essence become a team of monkeys. i do understand that monkeys can’t be experts. but then again neither are most students, work from home parents and “normal folk”. right?
can you really have it both ways?
either you have people that truly differentiate mahalo from wikipedia editors, which allows you to take on a google in the mano-y-mano battle between humans and algorithms… or you don’t. you’re just another “content” creator. you can’t speak of pagerank in the past tense (which you did) and then say that you’re bootstrapping the very same technology.
i mean, you can, but you won’t be taken seriously by people that have been building interactive systems since you were selling your SAR rag. sorry, but that’s the nut of this.
back to the problem with trying to link “expertise” with linking to a subset of links available to everyone. it’s completely subjective. are mahalo pages useful? sure. are they any better than wikipedia? aside from latching onto the zeitgeist term/result highway, nope. not really.
like i said in the post, all this is fine with me. make a bunch of bucks. trickle it down to your guides. fine. but CS was a conference about online creativity and my definition of creativity doesn’t quite extend into creative spinning.
i didn’t attack you. i called “human-powered search” out as a marketing scam. the same thing goes for the catch phrase, “the expert web.” i apologize. those are brilliant buzzwords.
btw, it’s funny you mention the whole “blogger” thing. i shut down my personal blog a bunch of months ago because i got caught up in the intra-day addiction of pointing to things instead of creating things.
i became a fucking monkey.
i’d rather use the technology than it use me (which any intra-day blogger with a degree of apperception and honesty would agree happens). i’d rather extend my voice, my experience, my knowledge into a blog, rather than add bridge links to shit and reap the benefits of attention.
what i can’t understand from your last paragraph is:
are you conceding that your google / wikipedia browsing “experts” are actually link bloggers?
and if so, how was i attacking you in the first place? because you’re their… leader?
I tried Jason’s new toy on several occasions and found the Google based results at the bottom to be far better than anything his people had placed at the top. Mahalo doesn’t even compare to http://www.bloggingpoet411.com and I’m neither a technician, a web guru or a slick talking salesman.
it’s pretty bad.
check out the longtail test i ran. compare that edge query with the google results.
then check out one of mahalo’s news pages (turkey bombing kurds) and compare it with the depth of a wikipedia page. the mahalo page uses the same AP article three times. why?
because the human “experts” manage their own pages without the quality control of the wisdom of the crowd.
mahalo vs. google ain’t happening.
mahalo vs. wikipedia is a joke.