Earlier this year, Powerset made a lot of noise by securing the rights to natural language technology from Xerox Parc as it set off to index the Web and building a search destination site.
According to the much covered VentureBeat story:
The move is significant because Google’s own technology, based on “page rank,” has been virtually replicated by other search engines like Yahoo and MSN, and so isn’t as difficult to emulate as it was a few years ago. Powerset could possibly steal a lead if it improves search results by a significant measure with natural language and simultaneously incorporates a near-equivalent to Google’s existing capabilities. Powerset has been hiring lots of Yahoo search experts and others, to help it do that.
That led me to ponder: should Powerset and Podzinger (now Everyzing) merge?
Today ZDNet’s Dan Farber reports that Powerset is out of stealth mode and the company’s ambition is “boundary-less”:
“We after after a pretty big goal–replacing the core of the search engine, the patents and technology are locked down,” said Steve Newcomb, a Powerset co-founder and COO, which won’t be open to the community until September.
I am certainly a believer that search is still very much in its infancy technology-wise, but from a business standpoint, it’s almost a fait accompli that Google will be dominant in it, much like Microsoft became dominant in software (related post: is Google the Standard Oil, AT&T or Microsoft of the 21st century?).
The main reason we think Google has checkmate the search industry, of course, has a lot to do with:
- distribution trumping technology in search these days
- advertising being bundled into Google’s core search offering
Of course, it is true that what made Ask Jeeves so interesting and promising in the late 1990s was its promise to deliver on natural search. But it should also be said that the failure to deliver on such promise is exactly what made Ask Jeeves crash and burn. Ask dropped that strategy - and the Jeeves part in its brand - and became a run-of-the-mill, standard search engine. Today’s it’s part of InterActive Corp. after Barry Diller acquired it for $1.85B in January 2005. We encouraged Mr. Diller to drop the $100M ad campaign and acquire companies, such as Powerset, instead. Read that here.
In other words, reinforcing our theme that search is the hubris of mankind on the Web, Powerset better be careful about its claims, and lord knows it’s aiming high. Problem is, it’s a lot of buzz and hyperbole right now. Let’s count the buzzwords (in bold).
From the same ZDNet article:
- Powerset leverages the wisdom of the crowds for development.
- “Imagine a mashup between Facebook, Digg and Google Apps, but you get to participate in the building of the products that sit on top of our platform. You log into a social network, like you would Facebook, and you get certified to be a Powerlabber. Once certified you can join different interest groups, such as travel, and participate in idea and mashup competitions. QA is embedded and its all bloggable.”
I know what you’re thinking: “but Ash, these are not buzzwords, these are massive trends that Powerset is embracing and leveraging.” Read on, please.
“We want as many people in Powerlabs to help us build and test the product. Powerlabs tells us when we are ready to go. We could have 50,000 people QAing our product,” he added. So far Powerset has 10,000 Powerlabs users. “Imagine how many widgets that could sit inside of Facebook, MySpace and even Second Life. It gives us the ability to launch with an extremely passionate set of people,” Newcomb said.
“We have people who were interviewing at Google, and now they would work for nothing at Powerset to be part of the company,” Newcomb said. “The passion is unbelievable. Powerset is ripping the core out again, like Google did, and they just want to be a part of it.”
Oh-oh, that last sentence is eerie: we’re basically reinventing the wheel, by jumping on board of a number of things that were hot and in vogue in 2004 (blogging), 2005 (social networks) and 2006 (widgets). Sure, these things are more pertinent today than ever, but folks, let’s get a grip. Let’s actually see it in action!
Like my concerns with Jimmy Wales’ Wikia search project, some of the loose arrangement will hamper its growth:
Regarding intellectual property for those contributing to Powerset, there may or may not be an agreement with those participating in Powerlabs, Newcomb said.
Powerset hopes that building a platform and attracting developers will save it from the fate of other search startups.
I’ve said this frequently, part of the reason some companies succeed online is the speed and focus of a very small core of developers: Yahoo!? 2 people. Google? 2 people. Facebook? 1 guy. See a trend. The wisdom of the crowds stuff is great in theory but tends to dumb and slow things down in practice. Oh, and please don’t mention Wikipedia, which is to some extent copied and pasted material from elsewhere from the Web. That’s not wisdom of the crowds, in my humble opinion.
As I continue to read, all I see are red flags:
”We are trying to challenge everything out there,” Newcomb said. He said many search companies had a 90 percent attrition in attention not long after their announcement. “After a few blogstorms they die–it’s the dead cat bounce model. We want to launch with super passionate people–David versus Goliath people and with the people who can build products off of our platform.”
”We are trying to challenge everything out there”? You sure that’s wise. Last time I checked MSFT too was doing that: Sony and Nintendo via XBOX; Apple and Sandisk via Zune; Google via Live.com; Yahoo! via MSN.com… I commend it for not sitting on its software business cash hoard, but I’m still curious if the “challenge everything out there” mantra works. Yes, that’s a bit ironic coming from me given the range of properties and applications we have, but beneath the surface, most of that all fits together. I guess, of course, Powerset would argue, so does their strategy.
Which raises a question: what is the strategy?
“Google is Goliath in this battle. Screw alpha, beta and blogs to launch–[Powerlabs] is the way to do it. The power of the blogosphere and the people active in our community is a big force, and it is a big deal to take on Google,” Newcomb said. “We are not anti-Google, we just believe the next decade is about computational linguistics.” However, Newcomb said he does “want to own the space.”
Effectively, Powerset is, if I understand this correctly, running with Wales’ idea of employing the wisdom of the crowds to scale (not in the open source context, since “patents and rights are locked down”). I’m not sure how well this will succeed, but despite our apprehension, we do with them well.
“Powerlabs will tell us when we are ready to go, and ready to build out the index,” Newcomb said. “We are in a financial position to wait until it is ready to go, and not a crap product.”
“We are betting that we win, but we don’t know, but we do know that search is going to get better because of it,” Newcomb said.
If that last part turns out to be correct, then we commend them for trying. There are good aspects to the potential of Powerset, but as Nitin Karandikar points out at The Software Abstractions in his thorough overview, we’re still talking potential. Tech Dirt seems to echo pretty what I’ve covered here, too.
“We are betting on our index,” concludes Newcomb.
Yikes. I give them a good chance of building a good product, but on the business front, it’s an uphill struggle. My gut says there’s a lot of bravado here to get a “too good to turn down offer” from Yahoo!, Microsoft, InterActive Corp., Time Warner AOL or News Corp.
Disclaimer: Our sister site MetaMojo.com is a vertical search engine. It’s got a place online, but we don’t for one second think it will own the space or take on Google. It will own the Mojo Supreme space, though, and that’s really all we think is up for grabs these days in search: your own backyard…
Google is the king of search, not by doing anything revolutionary in search, but by simplifying search and improving peripheral sectors like advertising. It did that by acquiring Applied Semantics’ technology and buying its potential competitor Sprinks (see more on this in our Top 10 Best Web Acquisitions of All Time here). The lesson: to win in search, there is no need to radically revolutionize the space, you just need to improve something tangent and you just might win.
Powerset: Natural Language Search
There was a lot of noise last year when Powerset raised a whopping $12.5 million financing on a large valuation. It was not clear then to the average outsider or onlooker how Powerset - or anyone else, including our own MetaMojo.com vertical search unit, frankly - would try to beat Google at search. MSFT’s $7.5 billion R&D budget has yielded little, neither has Yahoo!’s search efforts.
Today Powerset came out and disclosed that it had secured a family of patents developed by XEROX’s PARC dating back 30 years. Naturally, the blogosphere lit up like a Christmas tree.
In a nutshell, Powerset focuses on deep, natural language search while Google focuses does not. My two cents, very simply, is that 80% of searchers and search queries are well served by the GOOG/YHOO/MSFT oligolpoly, and even if 20% could be better served by a search engine like Powerset, I doubt the masses would ever find out about, let alone migrate to a new player.
Need Balance To Succeed
You see, Google not only offers a search tool, it also comes with a built-in monetization component. This is the killer app, if we wish to use that term. That is also why companies that offer one or the other fail, as did Yahoo! when it did not have an organic search product, or MIVA today.
Podzinger: Video Search
All to say, when I read this news, I quickly thought of Podzinger. While Powerset is a startup, Podzinger is not a real startup, it is in fact launched by BBN, a 30-year old company, with a $5M grant. The company is looking at raising outside investment this year to, well, take on Google.
Podzinger is not a traditional search engine per se, in the sense that it’s not a text-based engine, it is a video search engine. BBN has in fact leveraged its speech recognition technology and immersed itself into the world of tagging and vertical niches to organize, categorize and retrieve rich media. Hmm… organize, categorize and retrieve, that sounds awfully akin to Google’s mission ”to organize the world’s information and make it universally accessible and useful.” In all fairness I don’t think Podzinger’s mission is to organize, categorize and retrieve rich media, it’s just what I walked away with when I spoke to President of BBN’s Delta Division and Podzinger CEO Alex Laats.
The company’s approach can be summed up by:
- Audio to Text Transcribing of Content
- Tagging
- Development of Vertical Categories
- Clustering of Content
Not only is Podzinger not a startup per se, but Laats is not your run of the mill startup executive, he is a seasoned technologist executive whose sold his former company to 3Com for $90M in 1999. From there he moved into venture capital, and that lead him to BBN.
Working in video content and search, I can appreciate what Podzinger is trying to do. I spoke to Laats and he outlined some of the challenges and issues facing online video and advertising: “A lot of the tagging in place is spam, it is not useful, to really organize the content properly, we need more metadata, and you cannot get by with titles, summary alone.”
Of course, the Web space has gone gaga over tagging, so hearing using tagging to address a problem is interesting. Laats adds: “what we do is a major part of the solution, add metadata to content. We add value to the value chain, which we see as a four-part process:
1. Creation of Programming
2. Access into Content
3. Syndication
4. Advertising
BBN’s bet through Podzinger, frankly, is that Google and Yahoo! do not attack rich media search, and so long as they omit that major part of the content pyramid, there is room for new players.
This is interesting, because Powerset believes that Google and other leaders in search do not take natural language search seriously. From Venturebeat:
Xerox PARC scientist, Ron Kaplan, in a back room when VentureBeat stopped by for an interview last year. Kaplan, who has led the “natural language” group for several years, joined Powerset as chief technology officer in July. This is a coup for Powerset, because Kaplan did not respond to some early probes from Google. In an interview, Kaplan said he didn’t believe Google took natural language seriously enough. “Deep analysis is not what they’re doing. Their orientation is toward shallow relevance, and they do it well.” Powerset, however, “is much deeper, much more exciting. It really is the whole kit and caboodle.” While natural language has been a vexing problem for decades, Kaplan said he believes it is ready for prime-time.
Are You Feeling Lucky?
Hmm. I do not think that Powerset or Podzinger alone stand a chance due to the distribution challenge they face. But together, it becomes interesting.
How come? Read on.
Let’s first take a step back to consider what Podzinger actually is, and what it is not:
Podzinger uses BBN’s speech to text (transcription) takes a speech signal and figures out the probability that a certain set of sounds match a known word in the dictionary. After that conversion is completed, then it’s regular text-search approach. The upside is that this is fast.
Another approach is that of Podscope. Podscope is better for non-dictionary names, such as names of people, or places one visits etc. This is not as fast as Podzinger, which is better suited for mass market search queries.
Podzinger believes that it is head and shoulders ahead of its numerous competitors. Technically, those vying for supremacy in video search are numerous. After all, we have a video meta-search unit (though that has more to do with us preferring to let others duke it out in terms of technological supremacy and focusing on creating a good user experience by leveraging the best video search techniques out there). The only competitor that Laats considers worthy is Blinkx, but he is quick to add: “go and test the two and judge for yourself.” Laats is quick to add that Podzinger’s success can be measured by how many new media (blogs) and old media (CBS Radio, Intercom) companies work with it.
A Solution for the Masses?
The reason why I see Powerset working with Podzinger is that Podzinger has what appears to be a good approach towards video content search: recognize speech, transcribe it, then index the content. That means that Podzinger could be a rleevant player in video search. But what about text-based content? Well, there Podzinger has no product, technically.
Powerset, on the other hand, might have a fine product for the technorati, but will the masses really care about the benefits of natural language search? Well, they might if it could be bundled with video search. Think about it, locating content online is akin to finding a needle in a haystack, finding video content is even more daunting.
RECOVERY VS. DISCOVERY
John Battelle referred to the nuance between “discovery” (finding something new) and “recovery” (finding your way back to something you found before and want to see again).
With text content online, I’d say, what 80% of it involves discovery, right?
With video content online, I’d say over time, 80% of the challenge involves recovery because the online-only original video content (like that on WatchMojo.com, for example) represents a small percentage of all of the video content out there. Sooner or later, if media becomes increasingly digital, then the bulk of video content should in theory be the stuff from TV, right?
Well, that depends on many things. We have yet to see all of the text based content from offline migrate online, so who knows… but we’re straying too far from my hypothesis right now.
My gut says that if Podzinger can index video content by transcribing everything and then indexing it all into one big index of text-based content, then Powerset can apply its natural language technology into the last mile and offer a better user experience than Google. Google then would be beaten in video and text based content and this would actually be somewhat more material.
Show Me the Money
But then and there, there is still the lack of a monetization issue. Hey, maybe MSFT or Yahoo! can buy these companies, they have paid search results, right? Or, maybe Google will buy them both. Who knows. Can you tell I have a fetish for M&A (if you cannot, just keep reading this blog over the next few weeks or month).
Climbing the Content Pyramid
What makes the entire monetization challenge even more glaring is that the company first tackled content from podcasts, then dived onto the YouTube phenomenon. In other words, it has gone after the lower or middle part of the content pyramid. The higher end of this pyramid being made up of online-only producers of video content and offline video content. I could be wrong, but the course to success to monetizing video content does not start with podcasts and user generated videos.
Laats clearly realizes that the bulk of YouTube’s content is “unmanageable mass due to low quality content and copyright issues,” but we have yet to see how the company plans of indexing the high end of the video content spectrum.
Can’t Hear Sound?
Of course, by now you should ask: what about the content with no audio? This is where the tagging enters the mix. By tagging and then effectively clustering the video content, Podzinger sees an advantage into categorizing the mish mash that is online video right now. Like many others, Podzinger intends on building channels of content to make the user experience more manageable.
We do not doubt that Podzinger’s approach to the content equation can win in video. But for Podzinger to become a household name, then it needs to address the non-video segment of search, hence our crazy association with Powerset. But don’t count on it.
More important, for Podzinger to be successful at making money, it needs to develop that elusive business model for video advertising. Laats mentions, like many others, that video advertising will be a mix of pre, post rolls, banners etc. All of the people in this camp might be right, Google was not revolutionary in search, it was evolutionay. Maybe the winner in online video advertising need not be revolutionary either, but rather, evolutionary.
One thing is for sure, text search ads, which Google won in, is results and performance oriented, whereas video advertising is brand messaging. This entails an entirely different outlook to succeed in. And maybe, the more I think about it, the idea of hooking up natural language search with transcription search is not a bad idea to beat out Mountain View.
PLAY
