Our predecessors in the video content industry’s transition to digital were recently quoted regarding barriers they are facing to monetization. I was happy to find they map directly to three areas ffwd is innovating. First the quotes:

•  ”the killer video app will be an intelligent piece of software that creates a mode-sensitive, personalized [program guides]” - Veoh CEO Steve Mitgang
• “The one thing [YouTube] is least good at is the social piece, where users create their own community around a shared interest,” imeem CMO Steve Jang
• “consumers…shouldn’t have to think about whether it’s IPTV, or online video, or cable or whatever. The experience needs to be the same” - Ben Huang, Product Management Director, Microsoft TV

And now our innovation areas and what we’ve rolled out so far

• connecting the videos - the adaptive sequencing engine
• connecting the audience - passive sharing
• connecting the venues - the Facebook and Wii applications

It’s good to know that when we launch, we’ll unlock value for the whole ecosystem.

Ray Ozzie rocks, no doubt about it, but he’s evidently not immune from the particularly virulent strain of “blind to the world” endemic to Redmond. Take this bit quoted at Techcrunch: “Just imagine the possibilities enabled by centralized configuration and personalization and remote control of all your devices from just about anywhere. Just imagine the convenience of unified data management…”

Just imagine? Those of us building cross platform web-aps live in that world right now! Our imaginations have long given way to implementation. Right now, you can program your TV via the web using the ffwd service and our first two client application prototypes (the PC client and the Wii client). Soon we’ll be adding a mobile client (preview your program choices in transit) and a social client (watch what your friends have or will watch). By the end of this year we plan to have a full featured API that will let developers create clients for ANY standards-based, Internet-enabled device.

He then goes on to talk about “…the transparent synchronization of files, folders, documents, and media. The bi-directional synchronization of arbitrary feeds of all kinds across your devices and the Web, a kind of universal file synch.

Synch? What’s there to sync when every device is using the same data repository? The the web-at-large multi-directionally handles its native “documents” (HTML), “folders” (XML feeds), “media” (Embeds) using the “file” system of URLs. Perhaps Microsoft (and now Ray’s) problem is that none of those web natives are created using Microsoft software. The simple solution is to include with Vista the equivalent of a dot-Mac account (which I would support if only because it may push Apple to do the same).

So what is so important that I make fun of Microsoft for being blind to it? In the world at large, the PC (personal computer) is dying on two fronts:

1. The primary copy of our new data is more likely to live on a public server that on a personal device (eg. Flickr, YouTube, there isn’t won’t be even a copy of Facebook data on a PC)
2. The computer is giving way to the Internet-enabled device. The biggest threat to Microsoft’s play for the living room isn’t Apple, it’s the TV manufacturers who are rigging their displays with just enough to get online (btw, this article deserves it’s own future post because it best describes the vision of the future ffwd enables).

I think he understood this even a little bit into his experience at Microsoft, but these new statements suggest either he since stopped paying attention, succumb to pressure to quit the Creative Commons nonsense and get on with productizing his ideas, or (and this is the most disturbing) he is building his vision with the PC (or worse, a specific OS) in mind.

Please Mr. Ozzie, for you and us, make this MESH thing conform to the values implied by this as yet unrefined and unofficial war cry of web application developers everywhere.

“give me a standards compliant browser running on good hardware and a light-weight OS and I’ll do the rest.”

Update April 2008: Microsoft is releasing Live Mesh to a small group,  estimating general beta before 2009

The controversial Wisdom of Crowds concept popularized by New Yorker columnist James Surowiecki has come under fresh criticism. In March  The Economist, Newsweek, Bryan Caplan (author of the Myth of the Rational Voter), and Danny Dover all came out against the idea. Here’s a few choice quotes:

The Econmist on Wikipedia: “there is a limit to how much information a group of predominantly non-specialist volunteers, armed with a search engine, can create and edit he Economist on Wikipedia”

Danny Dover on the the fickleness of the mob

Bryan Caplan at SxSW: “The miracle of aggregation fails and it fails very directly.”

Newsweek quoting Jason Calcanis: “The wisdom of the crowds has peaked”

These would be damning cases of failure…if they were relevant. In reality, they only illustrate how difficult it is to make a crowd in to a wise crowd because  crowds in general lack one of the following characteristics of a wise crowd: diversity of opinion, independence, decentralization, intelligent aggregation, emotional balance.  To wit, the economist article  is really a bunch of illustration of how Wikipedia  is too centralized and politicizes aggregation. Likewise Danny Dover and Bryan Caplan’s criticisms are on failures of emotion or stupid aggregation.

To put it another way, I disagree with both Jason’s premise and conclusion. Web 2.0 was not about the wisdom of crowds. It was about crowds genearlly (social networks), a few somewhat wise, but none fully and intentionally so. Therefore, I conclude that Web 3.0 is a race between the return of experts (Mahalo) and attempts to wrangle the web 2.0 crowds into their wise form (Delver).

I was surprised at how little the speakers and panelists in the Video Search Summit mentioned precision (the measure of how many [top] results match what the user is looking for). In search, precision is one of the key drivers of user satisfaction, and there was virtually nothing said about it. I suspect this is because there aren’t many great things to report on this end…and there won’t be for a while.

One of the rarely-discussed truths of text-box search is that every search has an unstated predicate in mind. The user types in something like “hdtv”, and the search engine implicitly chooses which questions about HDTV will be answered with the top result(s). 

Google beat AltaVista in the web search war because it answered those questions better. It did so by implicitly determining from link structure which question people usually asked for a given search term and serving that result first to everyone. For example, with “hdtv”, the question might be, “What are options online for buying an HDTV set?” AltaVista’s algorithm basically looked for the pages that mentioned “hdtv” the most times, which might be technical documents describing how HDTV works. More people care about buying one than building one. So Google wins. 

For every query, the unstated predicated will break down differently among the users. For some queries, the breakdown may be 98-2. For others, it will be 50-50. For others, it will be 40-20-10-5-3-1 (etc.). Google wins by picking the top one. Even if fewer people are happy with a Google search for the 40-20-10 one (where 40% of people are happy with the top result) than the 98-2 one (where 98% of people are happy with the top result), they still beat or tie the competition with the 40-20-10. How would the competition do better — by picking the 20? That only makes more people unhappy. 

Video search is harder than web search because on average, the breakdowns are more fragmented. Some player will by definition come out on top, but the users aren’t going to be as happy as they are with Google’s web search. In the whole summit, the only comment I heard about good precision was when CastTV’s Alex Vikati showed a side-by-side comparison of several engines’ search for “Beyonce”. Her point was that users looking for news of Beyonce’s recent marriage would find the right result easiest on CastTV. But this presupposes what most users are looking for is the marriage. For someone who wants to see her music videos, or just wants to gawk at her, that is the wrong result. On top of all this, Google’s link-structure approach won’t tell you that the right video for Beyonce is news-related, because it won’t be true until the week she’s in the news. 

Another, subtle disregard for precision was that the most oohs and ahhs over technology concerned those players starting to use speech-to-text and image recognition fancy technology, and useful in isolated cases, but the basic truth is that those technologies have inherently lower precision than matching textual cues, so they boost recall (which, with searches returning 450,000 hits, isn’t the problem in very many cases) at the expense of precision (which is already taking a beating). 

Good precision is inherently hard to achieve with video search, and however the top player performs, they’re unlikely to be a heck of a lot better than second, third, and fourth place. Video search is going to be a commodity, where there are lots of players and no breakaway winners and no raving fans in the userbase for any one service. Search is going to be so-so. Video discovery is the area where the play for high user satisfaction will take place. 

The common wisdom on ubiquity is that consumers want everything, everywhere…a raw ubiquity of power. While this is probably born out of championing consumer choice, the usefulness rapidly diminishes once the everything part passes a certain scale or the everywhere part succumbs to bloat. For instance, in a world where 600 minutes of video is uploaded to You Tube every 60 seconds, I don’t actually want everything (emerging term: choice fatigue). Similarly, the Playstation 3 is part of so many everywheres that it ends up nowhere. Nextflix by contrast thrives by offering a sensible ways to limit choices and relax the pressure of real-time decision making (just add it to the queue and forget about it until it arrives at your door). The Wii doesn’t play Blu-ray DVDs or store video or do calculus, which makes it easier to turn it on when I want to play a game.

On the other hand, I do wish I could add to my Netflix queue from my mobile phone (perhaps via text message) and that I could buy Wii store games from my laptop. Similarly, ffwd’s vision is that applications on different platforms should communicate information to each other but should focus on functions that conform to the highest and best use(s). This sort of integration implies a limited feature set in a given platform, but increased usefulness across all relevant devices. Additionally, we lack presumption about what the highest and best use for a given user is on a platform, but rather than pack all functions in to “cover bases”, we suggest making client applications configurable. This is a ubiquity of presence, the optimal experience is threaded through all the platforms that do their part in completing a task.

Yesterday at the Television of Tomorrow show, I demoed for the first time how ffwd enables asynchronous “two screen” applications. It’s unremarkable in a way: use your laptop to find web clips that you think you will like and queue them for your TV, then switch to a box that connects the TV to the Internet (our choice is the Wii) where with one click you can start a streaming sequence of those selections and one-button skip through the ones that turn out to have been bad choices. What is significant about our implemented is actually (and I realized this only as I was demoing it) what you can’t do on the TV client: you can’t search, you can’t browse user profiles, you can’t share videos. In theory the user could configure the client to do those things, but we chose not to because those functions are more easily performed on the laptop. This is both an uncommon technology and an uncommon way of thinking about technology.

There are many examples of how pragmatically difficult it is to simultaneously hold both these ubiquities. Microsoft may have an identity crisis because of it. Advertising creative types are coming to grips with it. The kids, well, they prefer the ubiquity of presence, and that’s really the prospect that matters. Interestingly, there is one group that’s been managing this transition for a while, and they are doing quite well at it: interactive TV programmers. The example of the moment is Lisa Hsia who keeps extending the presence of Project Runway and won an award today for the effort. After my presentation, I had the pleasure of meeting people who have 10 years of experience with interactive video applications and who learned the hard way that (among other things) a two screen (laptop/TV) application can be better than trying to stuff interactivity into the set-top box or TV on the laptop. It’s not uncommon for pioneers in a new medium to presume that the old medium failed because of the people, rather than the intrinsic limits of the medium, and they then proceed to fumble through the same early confusions. Looking to avoid this mistake, I’ll bring this lessons of interactive TV to bear on our API launch later this year.

Mark Cuban asks the rhetorical question, “Will you watch what I watch?” to probe a question that is central to ffwd’s vision: how will viewers decide what to watch in a world of limitless choices? He dispenses with search as not providing real access to limitless choice (functionally limiting choice to the first page of results). I would add that search engines only enter the flow after most of the decision process is over. You must already know well enough what you want to watch to formulate a search string. He further questions if the currently ascending model of the celebrity endorsement, illustrated by Oprah, is destined to be the king.

At ffwd, we have a systemic understanding of the landscape that may help Mark’s inquiry. There are two axes to consider. First the archetypal cultural role at work here is the curator, the custodian of a collection - the Who part (like a museum curator). Secondly, the fundamental cultural operation in question is the transmission of information, or how the curator documents the collection - the How part (museum curators publish The Catalog for this purpose). The product of the curator and the transmission in video traditionally (pre-computers) has been called programming. I really want to emphasize that the essential problem post-computers is still the same (say it again, programming). What has changed is the number and configurations possible for the cultural components, the curator and the transmission method, has expanded

Some examples of how this this categorization plays out:

• Who: Corporation, How: media empire, Example: HDNet
• Who: Individual Celebrity, How: media empire, Example: Oprah
• Who: Common Collective, How: Democratic media, Example: VideoSift, MeFeedia, VodPod, YouTube Homepage
• Who: Celebrity Collective, How: Democratic media, Example: Funny Or Die?
• Who: Artificial Intelligence, How: Recommendation Service, Example: Matchmine, Taboola
• Who: Artificial Intelligence, How: Internet Publishing: YouTube recommended videos
• Who Corporation, How: Internet publishing, Example: Next New Networks

Like the periodic table, a structure like this makes it easier to see the relationship between known quantities (like the two examples in Mark’s post, HDNet and Oprah) conceptualize another options by mixing and matching the variations of each sub component. For instance:

• Who: Individual Celebrity, How: Internet Publishing, Example: ??? Could this be where Mark Cuban himself fits in? or Robert Scoble?
• Who: Corporation, How: Democratic Media, Example ???
• Who: Artificial Intelligence, How: Media empire, Example: ???

Well, we are busy enough allowing for the integration of any programming method and finding ways to measure their relevance to a viewer. We’ll leave the speculation over and creation of these possibilities to others. Hope this helps, Mark!

This isn’t going to be another post comparing the merit of the Wii to the AppleTV. But it is important to state for the record why a team of Mac fans chose Nintendo’s hardware over Apple’s. Consider it a wish list for Mr. Jobs:

• standards compliant browser - our biggest unmet need? support for Flash 9, but that’s really Adobe’s ball and we can rectify that when we spec our own hardware/software stack
• support for third party developers - Alan Quatermain is a genius angel for making it possible to build applications in anticipation of some unspoken future when Apple lets third party developers at their box or at least promises not to firmware “upgrade” them off. Nintendo actually customized the Opera browser to allow functions that are particular to the Wii experience (like the Wiimote) which is a great help for developers.
• The Wimote - if you haven’t used it, you may not know what I mean, but I think the biggest interface advance last year was not the iPhone’s touch screen, it’s the Wiimote IR sensor. Did you know it can point at stuff and drag it around from like 10 feet away. Seriously, you’ve got to come over to our office and try it.
• number of devices connected to TVs - total Apple TV sales are measured in thousands, Wii sales are measured in millions (and Apple TV had a year head start).

Perhaps the biggest gap for Apple TV for our purposes is a cultural one: the overwhelming download mentality of the product design. iTunes and the iTunes Music Store were breakthroughs for the digital music transition: a) preserved the transaction business model b) downloads allow for easiest repeat usage. Problem is in video a) repeat views are the exception b) the business model is media based. Hence the initial linking of AppleTV to iTunes on a desktop, and the persistence iTunes download distribution even though Take 2 is freed from the desktop proxy is a square peg/round hole.

But it wasn’t until we became a Wii developers that we realized the unique advantage of the device. Even if the Apple TV fills all the gaps above and forgoes the download mentality, it will still lack the Wii’s spirit of approachability. The Wii is the most social hardware device I have ever used. Obviously, devices like the mobile phone are more pervasively revolutionary for social interaction so that’s not what I mean. The Wii is social in the same way IRC was: it empowers/emboldens shy people to participate in social activities. It breaks down real barriers to interaction and assumptions of power (but that’s a topic worthy of discussion all its own). Suffice an example: my 5 year old niece can legitimately beat her dad at Wii bowling, and he still finds enjoyment in the subtleties of the control set. The Wii’s inherent accessibility means it is a platform to reach people who otherwise wouldn’t care to participate/consume the still rarefied world of Internet TV.

This is all to say that we believe the potential disruptor here isn’t (primarily) our (or any) application, it’s the Wii itself. One thing worth noting: we are by far not the first to try Internet TV on the Wii. Sofa Tube gets credit for planting the idea initially in my head back in 2006. A year later, but perhaps more famously, Stumble Upon took a few weeks to make a Wii custom interface for Stumble Video. What we realized is in our few weeks of working on it is that if you play to the Wii’s strengths your application becomes implicitly friendlier. If ffwd takes credit for anything, it is in setting the upper most bar for a video user experience on the Wii. And what we accomplished truly shocked us.

We think “The Wii is a social entertainment device” is an important realization to share, so for the first time we are actively encouraging the spread of a meme. We’ve got one 2000 Wii points card (that’s approximately 2 classic games) for up to 5 bloggers who trackback this story from a blog with a page rank of at least 5 (as proofed at PRChecker). If we don’t get enough over 5s we’ll go to the 4s and then the 3s (yay! everyone can play). If this is an effective incentive we may go buy more cards to give a way. First come first served and keep in mind we’ll eventually need your contact info to send you the card.

Thanks to Ian Schafer the CEO of Deep Focus, the web got a rundown on YouTube’s Videocracy event in New York last night. The audience was apparently mostly marketing/advertising folks and the message was geared toward them, but there were a few product developments mentioned that suggest YouTube is about to go through the same identity crisis that the early players in the digital music transition went through.

The new features announced amount to early steps in the “socialization” of YouTube’s distribution hegemony. I could be making the wrong analogy, but if it is anything like music, the essential strategic logic is that now that we appear to be on the road to lock-in on distribution/audience share, what value can we add on top of that to increase margins. This is coupled with trying to curb the low level of engagement (time spent per session, for instance). This line of thinking almost always results in some form of “power to the people” tactics for promoting content. To wit in YouTube’s case:

• What YouTube tries to differentiate by calling it “active sharing” but is actually yet another implementation of the status functionality made popular by Facebook
• Collaborative filtering (again)
• Simple audience management tools as a preface to asking (soon begging) content creators to make YouTube their publishing homepage (i.e. don’t buy your own URL and identity)
• ubiquity, meaning publish once watch anywhere

This is the most reasonable direction for YouTube to go, considering their options, but as the title of the event makes clear, it is really an attempt to claim the mantle of an idea (video democracy) which, and this is my key point, they are intrinsically against.

Full democratization of a space/industry should include a choice of distribution options. YouTube’s conceit over the next year will be: anybody can participate, everybody has opportunity, and the vox populi will be heard, choices are limitless…as long as we all to choose to put our very colorfully differentiated eggs in YouTube’s distribution basket. I repeat, this is a reasonable claim for YouTube to make and I’ll add a reasonable bargain for the creator community to take, but with one important caveat: that there remains no other option.

What YouTube will (should, must) try to do in the video marketplace is exactly analogous what Amazon has tried to do with Marketplace, Fulfillment by Amazon, Advantage, etc. Amazon attempted to leverage their retail hegemony to co-opt the independence of smaller sellers. Likewise, YouTube is leveraging its distribution hegemony to co-opt the independence of smaller publishers. However, Amazon is not the company synonymous with the democratic marketplace. eBay is.

eBay was founded a year after Amazon, and while it wasn’t exactly conceived along this line of thought, it does represent the best answer to the question: what would a place to buy things look like if you even democratized the sales/distribution chain? There would be no hegemonic warehouse, no single shipment provider, no single payment method (to wit, eBay’s failed attempt to push BillPay). eBay, was originally thought of just as an auction company, then a marketplace, but in the context of this discussion I’d like to propose they are, in fact, a transaction information organization company.

One obvious criticism to this analogy that I’d like to put to rest quickly is that we shouldn’t compare a retail space to an ad supported one. That criticism is just currency semantics. While consumers don’t spend dollars at YouTube, they do spend time. In fact, this provides an opportunity to deepen the analogy. Just as it is in Amazon’s interest that you spend your dollars at their site on items where they have the highest margin (lowest cost). It is in YouTube’s interest that you spend your time at their site in a way that taxes their resources least (either watching video that someone else is paying to distribute/promote, or click around not watching video at all). eBay on the other hand, and this is a restatement of my key point doesn’t need to care about any of this…they care about the total number of transactions.

One last point, before I get to the conclusion you can guess is coming. There is no strong evidence that one of these paths trumps the other. eBay’s market cap is currently ~15% higher than Amazon’s, but I’m sure the opposite was true at some point. Regardless both are worth over $30 billion Both continue to suffer from constant pressure on their core businesses and the need to expand into other areas to justify the valuation (Skype, Kindle). Both companies have made missteps trying to adopt the other’s intrinsic natures (zShops, BillPay).

And now the proclamation you’ve all been waiting for. It’s time for an eBay of the video space and ffwd is committed to building it. A democratic video information organization company that doesn’t care which distribution platform you use and whose goal is to maximize the use of a viewer’s time (currency)? Call us crazy, but we won’t stop until we’ve either produced a worthy counterpart to hegemony, or flame out trying.

Get your thumb ready! We’ve been talking about this for a while, but I hope you can still share my excitement for getting it out the door. People’s eyes light up when I tell them about the future we are aiming for, channels of Internet content streamed to your TV, but then they come back to earth saying, but that’s not happening anytime soon. That’s what we thought, too: 2010 probably, 2009 maybe. but then last fall I was approached by a consumer electronics manufacturer with a question.

Q: What do you think Internet TV should look like?

A: Why do you ask?

Q: Well we are putting this here Internet jack in our TVs next year and we’re wondering what else we could do with it.

A: Next Year! Uh I mean, we’ve got some ideas. Interested?

Well now it is…next year…and I’m proud to reveal the first product to use the ffwd platform for delivering Internet TV to your, uh, TV. Everything you’ve saved to your ffwd account will be there waiting for you. You’ll also be able to discover new favorites based on your profile and what others are watching. Existing beta users point your Wii browsers to http://www.ffwd.com and we’ll automatically detect your Wii-ness. Others can see screen captures and more information at http://www.ffwd.com/wii

Happy channel surfing…ffwd can make it easier…but only your couch can make it this comfortahhhhhh…

At ffwd we believe that  over the next six years or so, some really smart people are going to develop really smart methods for recommending video content and our goal is to provide a frame work for those methods to get sorted out according to the viewers they work for. In the meantime, we are building our own recommendation methods because we need them to test the framework, but we  already discussing integration strategies with some early pure-play recommendation companies and working with Mashery on a system for user initiated private info sharing.It’s not obvious from the ffwd interface, so it is worth mentioning explicitly, as a video preference repository, ffwd wants users to be able to what they want with their data and where they want to. Therefore, at a very basic level, Michel can put away his worries that ffwd’s success will aggravate the zero-sum game effect. Quite the contrary ffwd benefit from the adoption of portable user profiles as he describes it.

The user gets to experiment with which site has the algorithms and user base to provide good ratings in which situations, and the services get to compete on how good they do their jobs

See we take this notion a step further, namely a user shouldn’t be forced to jump from site to site (UI to UI) in order to try out recommendation algorithms. Moreover, once they’ve found some they like they should be able to access them separately or in aggregate from a single interface. Most importantly they should be provided accurate info on how those methods are performing for them so they can guided amongst them wisely. Serving this set of needs is really where ffwd fits into the ecosystem. We plan to build a platform where a suite of general services are provided to recommendation companies so they can focus on building the best recommendation methods.