Fortune has a great story about the rising importance of recommendation engines and the search for the perfect one. Excerpt:
The Web is leaving the era of search and entering one of discovery. What's the difference? Search is what you do when you're looking for something. Discovery is when something wonderful that you didn't know existed, or didn't know how to ask for, finds you. When it comes to search, there's a clear winner - a $145 billion company called Google (Charts).
But there is no go-to discovery engine - yet. Building a personalized discovery mechanism will mean tapping into all the manners of expression, categorization, and opinions that exist on the Web today. It's no easy feat, but if a company can pull it off and make the formula portable so it works on your mobile phone - well, such a tool could change not just marketing, but all of commerce.
It has in-depth examples from Pandora, Slide, and Whattorent.com. It also offers this bit of speculation about Google's plans to get into the recommendation game:
Steve Johnson, CEO of Boston-based recommender system ChoiceStream, which provides the collaborative filtering engine behind AOL.com, Blockbuster.com, iTunes and Directv.com, says he's been talking to Google about building a system for YouTube.
As usual, TechDirt is not impressed.
You can do what Google Video and YouTube do—match the terms to the words surrounding a particular video such as the title, description, tags and other so-called metadata.
Or you can do what Dabble does—search all the text that is available, but then rank the results based on how humans interact with a particular video. These “gestures,” as Dabble CEO Mary Hodder likes to call them, include whether anyone on Dabble (or sometimes the Web at large) has blogged about a video, tagged it, linked to it, bookmarked it, added it to a playlist, or searched for it. I recently caught up with Hodder for an update on Dabbe;. She says:
The metadata is a mess. The first thing we are doing is making sense of the data. The second piece is that Dabble is working differently to perform search or understand media than other services. Dabble is both a social community and video search.
Dabble searches the Web for video, targeting more than 300 video hosting sites like YouTube, Revver, Metacafe, and Blip.TV, as well as video blogs and other sources. (Google Video, by contrast, does not actually search the Web for videos—it only searches videos that have been uploaded to Google Video). So far, Dabble has indexed close to 4.5 million videos. It is adding 80,000 to 100,000 videos to its search index every day (with about 65,000 of those coming from YouTube). "I think we are getting about half the videos out there to get," estimates Hodder.
That is not as easy as it sounds. For example, type in "crush or flush" into Dabble, and up comes a video I made and blogged about that shows a demo of that product. But it doesn't show up on Blinxk or Google Video. Why is this? Video search is new, and many sites make it hard for others to search their video. Spidering the Web, like Google and Yahoo do for regular Web search, doesn't always work.
The blogoshere is abuzz with a leaked memo written in October by Yahoo senior VP Brad Garlinghouse dubbed the Peanut Butter Manifesto. In it he warns that Yahoo is spread too thin and is all over the place. Excerpt:
We lack a focused, cohesive vision for our company. We want to do everything and be everything -- to everyone. We are scared to be left out. We are reactive instead of charting an unwavering course. We are separated into silos that far too frequently don't talk to each other. And when we do talk, it isn't to collaborate on a clearly focused strategy, but rather to argue and fight about ownership, strategies and tactics.
Our inclination and proclivity to repeatedly hire leaders from outside the company results in disparate visions of what winning looks like -- rather than a leadership team rallying around a single cohesive strategy.
I've heard our strategy described as spreading peanut butter across the myriad opportunities that continue to evolve in the online world. The result: a thin layer of investment spread across everything we do and thus we focus on nothing in particular.
I hate peanut butter. We all should.
We lack clarity of ownership and accountability. The most painful manifestation of this is the massive redundancy that exists throughout the organization.
Universal Music made good on its promise to sue MySpace, filing a lawsuit yesterday claiming copyright infringement by MySpace members. The two companies had been in negotiations to enter a licensing deal similar to the one Universal had struck earlier with YouTube, but Universal apparently also wanted to be paid for past infringements and the number MySpace put on the table wasnot high enough. That's when Universal and reached for its legal stick.
Is this just another negotiating tactic, or is Universal willing to take this copyfight to the courts?
One person, Fox Interactive Media's Ross Levinsohn (the News Corp. exec originally responsible for buying MySpace), isn't sticking around to find out.
The promise of Web-based enterprise software has been a long time coming. But for the past few years the only real standout in this category has been Salesforce.com, which now has more than half a million paying corporate subscribers and is on track to hit nearly $500 million in revenues this year.
But lately I’ve noticed that some newcomers are gaining traction delivering Web 2.0 software to corporations. These include Zimbra, Success Factors, and Rearden Commerce (which just got a $22.5 million investment from American Express).
Zimbra (an original Next Net company) now has about 4.5 million individual paying subscribers using its Web-based corporate e-mail, contact, and calendar software. And it only just launched last February. At a bare minimum, Zimbra is getting $1/user/month, CEO Satish Dhamaraj tells me, noting that the vast majority of his subscribers come from bulk deals through ISPs, and for corporate customers he gets $28/user/month or more. So Zimbra is making at least $4.5 million per month (or about $50 million a year) in revenues. Dhamaraj told me last week that his startup is not yet profitable, but it is “pretty close.” He added: “We are growing more than 100 percent in both revenues and customers, quarter over quarter.” That's sequentially. “Next quarter we are going to double,” he affirmed. (Dhamaraj didn’t show me any audited financial statements to back up this claim, so I am taking him at his word).
In the last issue of B2.0, I wrote about a startup called Critical Mention that is doing some amazing things with video search (through CriticalTV) and syndicating clips (through ClipSyndicate) of local TV news on the Web. Excerpt:
By indexing the closed captions for all the video stored on CriticalTV, the service creates searchable transcripts on the fly.
And the service is still getting smarter. Using speech-to-text software recently licensed from IBM Research, CriticalTV will soon monitor video that isn't closed-captioned and also search Arabic-language TV and translate it into English. Customers ranging from corporate PR officers to FBI analysts to oil executives will continue to be alerted via e-mail (with a link to the video clip) minutes after a company, product, or terrorist name is mentioned on TV here or abroad.
But while CriticalTV pays the bills, [CEO Sean] Morgan's ambitions go beyond mere monitoring. In an era when TV stations are losing their audience to channel surfing, commercial skipping, and the Web, the ClipSyndicate site promises to find a better audience for news broadcasts. Morgan has already struck deals with more than 65 affiliates and is in discussions with all the major station groups to host clips from their local news and other shows.
Once the video is on ClipSyndicate, publishers can search for segments, splice them, and stream them through their own sites. In effect, Morgan is turning one-time TV broadcasts into media chunks that can be remixed for more exacting online audiences.
ClipSyndicate will insert 15-second ads in front of each video stream and split revenue 50/30/20 with the affiliate and the web site, a model similar to NBC's new business, the National Broadband Co.
AOL CEO Jonathan Miller was replaced yesterday by NBC exec Randy Falco. That may seem like an odd reward for the guy who helped turn around AOL and shift it's strategy from relying on subscriptions to advertising. Partly, this shows how big a bet Time Warner (which is also my employer) is putting on Web video. (Did I mention this was going to be the year of Internet video?) Partly, there might be some truth to the speculation that Time Warner CEO-in-waiting Jeff Bewkes wants to have his own man at the helm of AOL.
Update: Now it looks like Miller protogé Jason Calacanis is also leaving the building.