Productivity - The Big Challenge For Mahalo
After watching a pretty interesting (and revealing) interview that Jason Calacanis gave about Mahalo, it got me thinking…
Jason describes Mahalo as human-powered search. That is, rather than coming at search from the “technology angle”, like Google does, Mahalo comes at it from more of an “editorial angle”. Well… that might be how it is right now, but I’m not sure that’s going to be how it ends up. To succeed, I think it needs a pretty major technology angle too.
What will that technology be? Well, the big challenge for the company will be the productivity of its human guides; and the company will need to develop technology that helps the guides be ultra-productive at both creating SERPS and also keeping them up to date. So, it will be the technology as well as people that are the differentiators for this business.
How can I be so sure? Well, one of my companies specializes in helping people be super-productive through use of technology, and I’ve seen this situation in companies operating manual business processes many, many times before. That company operates in a different domain - discovery of new drugs in biotech and pharmaceuticals - but the situation is essentially identical. In the case of modern drug discovery, a laboratory scientist (with the help of robot helpers) can , in a single afternoon generate so much new data, that it would take them months to analyse if they had to do the analysis without the help of advanced computer systems.
The point is this: anytime you have people involved in critical processes that involve large amounts of data and information as an “input”, those people quickly become overwhelmed if they don’t have sophisticated software technology software that eliminates the major bottlenecks. And by “overwhelmed”, I don’t just mean to the point of it being a little painful for people. Rather, I mean to the point where the entire business becomes unsustainable.
Actually, I suspect that the Mahalo guides are already starting to become a little overwhelmed, and they’ve only just got started. Let’s just take one example. The most popular search term on Mahalo at the moment, after the Mahalo FAQ, is iPhone. Yesterday, there was some pretty interesting news about iPhone that people searching the Web for iPhone would want to know. That is, it was announced that iPhone will be in the shops on June 29. But was the iPhone search results page updated to reflect that? No (at least, not that I can see). Now, this isn’t a criticism of the Mahalo guides, they are clearly busy creating new pages for new search results - and they can’t be expected to keep on top of everything that’s relevant. Instead, I would assert they should be using technology to help them do that - and specifically, unique technology that will help differentiate their business and create genuine competitive advantage.
Specifically, I think that what Mahalo needs is advanced software technology that can automatically monitor the Web for potentially interesting “chatter” about their key “top 25%” search terms. These systems should not only monitor this chatter, but also automatically prioritise it for human screening, and assign the screening to the most appropriate guide (based on expertise and current workload). Those kind of integrated “unstructured data” mining and screening systems are pretty sophisticated, and would take a while to develop. But without them, I just don’t see how Mahalo can succeed…
Asam Bashir wrote:
Strikes me that the kind of software technology that’s needed for this would also be the kind of AI technology the security services would be interested in and are probably already being developed. However, whilst this indeed would be clever technology to develop, I don’t think Mahalo as a concept will work. It seems to be yet another business plan that basically bottles down to trying to create a lot of page hits and views, but without any of the addicitive features of social networking sites. I’ve only had a quick look at the Mahalo site but the topics seem to be more suited to established portals where you can clearly establish authority and trust. Why would I trust a Mahalo guide in say a certain health topic rather then say WebMD. Sure, Mahalo might give me the latest papers for a medical condition, but this isn’t the information the public would want or need. In summary:
1. Why should I trust Mahalo in the absence of a clear mechanism to establish author/editorial authority in specialist subjects?
2. What’s going to make me come back, where’s the addicitive element?
3. Where’s the AI technology needed for the kind of advanced monitoring which would be needed without an army of guides?
Posted 05 Jun 2007 at 2:32 pm ¶
simon wrote:
Yes - the Mahalo business model is “lots of page views”. I think that’s fair enough, even without social networking. After all, Google gets lots of page views
For sure, intelligence agencies already have an aspect of the kind of software I’m suggestng. But it’s not that great, I don’t think. For a while now, various intelligence agencies have had, as part of their mission, a goal to get some of their technology out into the private sector. Whenever they’ve shown it to me, I have to say that I’ve not been that impressed.
To address your question 1. It’s a good question. Remember, though, they’re not actually creating content;they’re just linking out to it. I think the idea is that, over time, the community will help ensure they link out to the most authoratitive sites for a topic… they’re betting this will be better than Google.
2. What will make you come back? This is a really good question. Jason’s answer seems to effectively be “why *wouldn’t* you come back?” However, most people use just one search engine - Google. Clearly, though, a search engine isn’t really that sticky. The only thing that will make you come back is if it gives you fabulous search results. The bottom line is: Mahalo has to do significantly better than Google for the “top 25%” of web searches, otherwise they’re out of business.
Your point 3. That’s another good one. I think it would be possible to do a pretty good Mahalo-type “top 25% search term pre-computed results site” that didn’t require human guides at all, by using advanced AI software. It could have some of the same advantages that Mahalo should have i.e. virtually impossible to game/spam by using SEO. And, it could be much more up-to-date than the biggest army of human guides could manage. I suspect you could build in quite a lot of expertise into the “virtual” guides as well. Hmmmmm… that’s actually quite an interesting idea… maybe more than “quite interesting” in fact…
Posted 05 Jun 2007 at 3:17 pm ¶
Jason Calacanis wrote:
Mahalo for the feedback!
I think you make some great point about productivity. With regard to the iPhone SeRP (search engine result page) it has a RSS feed pulled from Google News on iPhone that keeps the page up-to-date. RSS is a big solution to keeping the pages up to date.
Also, if you look in the right hand column you’ll see the top submitted links by users… clicking it take you to this list:
http://www.mahalo.com/Special:Mahalouserlinks?page_id=7451
our hope is that the public keeps the site up to date and does half the work for us AFTER we have created a SeRP.
Mahalo for visiting Mahalo!
jason
Posted 05 Jun 2007 at 5:11 pm ¶