At the NMK Forum in London, Jason Calacanis has just announced Mahalo Greenhouse, part of the recently launched Mahalo human-assisted search directory. The Greenhouse will allow the public to add search results and, if accepted by the site’s guides, get paid for them.
Mahalo launched on 30 May at the D5 conference. It’s been billed as a human-powered search engine, but it’s more of a ‘human-powered wiki’ listing search topics and links instead of encyclopedia entries like Wikipedia. As a matter of fact, Wired called it “a version of Wikipedia with advertisements“. The launch was met with much fanfare and a fair number of questions. Would it scale? Could it beat Google and its voracious algorithms? Why would it work better than Ask Jeeves or ChaCha?
What would you do next if you were CEO of Mahalo? … Wondering if you guys were me, what would you do next with Mahalo.
In response to the suggestions, they will now allow the public to submit search results on the Mahalo Greenhouse site to be evaluated by the full-time guides that site employs. Right now, they’ve got 40 full-time guides, but they expect that to increase eventually to 100.
Jason’s thinking is that Reuters, AP and DowJones employ hundreds of people to write editorial content, why not employ 100 people to curate search.
He’s not trying to compete with Google and Yahoo on ‘long-tail search’ but rather focus on curated results for the top one-third of search. This is not about being broad and deep but about being relevant and providing results for the most lucrative search terms. Right now, they have about 5,000 search terms, but they plan to eventually reach 25,000.
To scale to that number faster, they decided to use the Greenhouse to crowdsource the best links, paying $10 to $15 per accepted submission. The more submissions a part-time guide submits, the more money they make per submission.
But what is the business model, and how will it scale to paying all of those guides? The business model is advertising. Search ads are the most desired ads for a reason, Calacanis said because people are indicating an interest at a particular point in time. They have entered something that they are looking for in a search box. It is not passive and wasteful, he said, like display advertising at a bus stop or on a billboard.
In many ways, it mirrors a commercial Wikipedia. Calacanis and Wikipedia founder Jimmy Wales have clashed over whether to include advertising to support Wikipedia, and it wouldn’t be unfair to say that Calacanis is out to prove a point with Mahalo. In a not unsubtle rebuttal to Wales, Calacanis said that if part-time guides don’t want their payment because they believe in the concept of free culture, they can donate their payment to Wikipedia. Mahalo has already earmarked up to $250,000 this year to donate to Wikipedia in lieu of pay for guides who request it.
Mahalo may sell their own ads in one to two years’ time, Calacanis said, but right now, he believes it would be a waste of time and money trying to sell ads on the site before it reaches critical mass, a lesson he learned at AOL.
But he knows that this will take time to build out the number of search terms and also to build the traffic necessary to attract advertising. “This is a big project like building the Brooklyn Bridge or Central Park,” he said, adding that he’s committed to the long term and has enough money to fund the site for five to six years. Backers include, Sequoia Capital (where Calacanis is an entrepreneur in action), Mark Cuban, Ted Leonsis of AOL, CBS, News Corp and Elon Musk, founder of PayPal and SpaceX.
Calacanis believes that Mahalo is needed because:
the internet faces an environmental crisis of spam, (search engine optimisation), phishing, adware and spyware.
The internet is becoming polluted, he said. The amount of bad information as well as good information is exploding. The internet needs curation. If nothing is done, he worries that in a few years the internet will become too difficult to navigate.
Some have questioned the value of subjective choices made by the Mahalo guides. Calacanis responds:
I would rather have a little bit of bias and debate and refine rather than have the machine get it wrong and not to get to talk back to the machine
Also, on controversial issues such as abortion or George W Bush, the site will list general information but also provide search results showing different points of view, such as pro and con, for and against.
Will they have Digg-like voting? No. Calacanis says that voting is meaningless because people often vote before visiting a site in Digg not after they have gone to the site to evaluate it.
Calacanis also wants to build in accountability for search terms submitted. People should own their words. Part time guides will have to use their real names if they want to be paid. In general, he believes that anonymity is useful in limited cases such as whistle-blowers.
Jimmy Wales is also working on a new search project possibly with some human element. The details remain vague, although Danny Sullivan of Search Engine Land has a good interview with Wales about the project from last December. Like Wikipedia, many expect to have some human element in the search project, but all Wales said was:
Exactly how people can be involved is not yet certain. If I had to speculate about it, I would say it’s several of those things, not just community involved with rating URLs but also community rating for whole web sites, what to include or not to include and also the whole algorithm … That’s a human type process that we can empower people to guide the spider
Calacanis doesn’t see Wales’ project as a competitor. “I know that will be a disappointment for the media. It’s not a battle royale.”
In the interest of disclosure, I conducted this interview the day before the NMK forum in my role as blogs editor at the Guardian, not simply as an independent blogger. In as much, I agreed to abide by an embargo this post until the announcement at NMK.