Posts for 'Video Search'

continue reading

Jinni, whose "taste and mood" video discovery technology allows viewers to get personalized TV and movie recommendations, has announced a license deal with Microsoft. Though Microsoft didn't disclose how specifically it would use Jinni, the company has been broadening its Xbox gaming platform for video entertainment, so it would fit well there. Jinni could also be integrated across Microsoft's online and mobile properties and in different devices like the Zune.

continue reading

continue reading

continue reading

continue reading

continue reading

EveryZing, the search and publishing technology firm, is announcing this morning that it has been chosen by FOXNews.com and FOXBusiness.com to power universal search for both sites. The deal means that current and archived videos, podcasts, articles and images on each site will be indexed and presented online using EveryZing's SaaS-based Universal Search Solution. The two FOX implementations are great examples of how EveryZing can cohesively present various media formats to benefit both the user and content provider. Tom Wilde, EveryZing's CEO walked me through the FOX implementations last Friday.

The starting point for content providers working with EveryZing is to have their content indexed, transcribed and tagged by the EveryZing system. For the FOX sites that meant millions of content objects, EveryZing's largest implementation to date. From the user's standpoint, the most compelling thing about EveryZing is the control and flexibility it allows to pull out of the index just the results desired and in the preferred media format.

For example, if you start a search with "Stimulus" you're presented with results ordered by relevancy. But if you select to filter by video, then you see just videos tied to the topic. Each video is presented with time stamps you can roll over to see the sentence in which the search term was used. Clicking on that time stamp takes you to that specific point in the video. Other time stamps are presented in the video clip as well, for easy jumping.

Conversely, if you're interested in a comprehensive package of all results tied to the keyword, EveryZing offers related "universal topic pages." So for "stimulus," the two related topic pages are "Stimulus Package" and "Economic Stimulus." Click on either and you'll see all results for these terms. A topic page is EveryZing's way of grouping all related assets onto one page, which enhances discoverability by search engines and engagement by users. On the "Stimulus Package" topic page, you can drill down by media type (e.g. video, story, blogs). You're also presented with a dynamically-upated list of related topics. For the two FOX sites, EveryZing has created 3,500 topic pages, along with 125,000 video landing pages. EveryZing also enables promotions of specific on-air shows that are related to the topic, a great tool for boosting visibility and audience.

With EveryZing's SaaS approach, the FOX sites are not hosting any EveryZing software. Instead, FOX has created the search results page templates, and when a user runs a search, the results are published by EveryZing into these templates and served (along with the videos themselves) by Akamai, which is FOX's CDN. EveryZing's model is to be paid a monthly fee on the basis of how much content it indexes and how many hits to the database are generated. All activity should result in another ad opportunity for the content provider, so as long as the content provider can sell its ad inventory, the model should be positive.

I've been bullish on EveryZing for sometime (see here and here) because it exposes content providers' burgeoning volume of video content to their users' well-established search behavior patterns. Importantly, by blending video with other media formats, EveryZing allows users to decide what format they want to engage with at that particular time. Because no two user experiences are ever the same and more and more content providers are utilizing different media formats, I see EveryZing's approach only increasing in value.

What do you think? Post a comment now.

Pixsy, a white label video search provider made an interesting announcement yesterday about the launch of its new "Premium Feed" service, which I think is another example of the Syndicated Video Economy that I've been talking about for a while now. I talked to Pixsy CEO Chase Norlin about Premium Feed to learn more.

For those of you not familiar with Pixsy, it has been quietly building one of the largest video indexes since its founding in 2005. To date it has mainly focused on licensing the index to partner sites which wanted to offer easy video discovery to their users. As more content providers have offered embedding, Pixsy also enabled found videos to be played right on its partners' sites. Even though activity has grown well, Chase is pretty candid about monetization to date being difficult.

Premium Feed takes embedding to the next level by creating a subset of Pixsy's video index that is both higher-than-average quality and has accompanying pre-roll and overlay ads. Then Pixsy is developing an economic relationship between the content provider and its publisher network by signing redistribution and revenue-sharing deals with both. Chase says that to date the publisher network has 45 million unique visitors/mo and that 1-2 million videos are in the Premium Feed.

One of those publishers is EgoTV, and I chatted with founder/president Jimmy Hutcheson to find out how they're implementing Premium Feed. If you look in the lower right corner of their home page you'll see 3 new "channels," Ego Cars, Ego Comedy and Ego Travel. Each of these are constructed solely of Pixsy Premium Feed videos that are curated by an EgoTV editor. In another example at Ego People, the 300x250 ad in the right column is now populated with the Premium Feed. This is a simple "highest-and-best-use" real estate decision: Jimmy explained that Premium Feed is yielding 2-4x as much net revenue for EgoTV as it would receive if it sold rich media ads in this position.

Yet each of these initiatives has its own somewhat differentiated value proposition and underlying technology approach. As syndication grows in importance, sites with strong traffic and an interest in incorporating video will have many choices. As to how they'll decide, Chase makes a good point: simplicity and one-stop shopping are always valued by resource-constrained sites. Providers that can address as many of these sites' potential needs will be in a strong position.

What do you think? Post a comment now.

EveryZing, a company I wrote about last February, is announcing the launch of its MetaPlayer today and that DallasCowboys.com is the first customer to implement it. My initial take is that MetaPlayer should have strong appeal in the market, and could well shake things up for other broadband technology companies and for content providers. Last week I spoke to EveryZing's CEO Tom Wilde to learn more about the product.

MetaPlayer is interesting for at least three reasons: (1) it drives EveryZing's video search and SEO capabilities inside the videos themselves, (2) it provides deeper engagement opportunities than typically found in other video player environments and (3) it enables content providers to dramatically expand their video catalogs, while maintaining branding and editorial integrity.

To date EveryZing's customers have used its speech-to-text engine to create metadata for their sites' videos, which are then grouped into SEO-friendly "topical pages" that users are directed to when entering terms into the sites' search box. Speech-to-text and other automated metadata generating techniques from companies like Digitalsmiths are becoming increasingly popular as content providers continue to recognize the value of robust metadata.

MetaPlayer takes metadata usage a step further by creating virtual clips based on specified terms, which are exposed to the user. A user's search produces an index of these virtual clips, which can be navigated through time-stamped cue points, transcript review, and thumbnail scenes (see below for example). The virtual clip approach is comparable in some ways to what Gotuit has been doing and is pretty powerful stuff, as it lets the user jump to desired points, thus avoiding wasted viewing time (e.g. just showing the moments when "Tony Romo" is spoken)

Next, MetaPlayer enables deeper engagement with available video. Yesterday, in "Broadband Video Needs to Become More Engaging," I talked about how the importance of engagement to both consumers and content providers. MetaPlayer is a move in this direction as it allows intuitive clipping, sharing and commenting of a specific video clip within MetaPlayer. Example: you can easily send friends just the clips of Romo's touchdown passes along with your comments on each.

Last, and possibly most interesting from a syndication perspective, MetaPlayer allows content providers to dramatically expand their video offerings through the use of what's known as "chromeless" video players. I was first introduced to the chromeless approach by Metacafe's Eyal Hertzog last summer. It basically allows the content provider to maintain elements of the underlying video player, such as its ability to enforce a video's business policies (ad tags, syndication rules, etc.), while allowing new features to be overlayed (customized look-and-feel, consistent player controls, etc.).

MetaPlayer takes advantage of chromeless APIs available now from companies like Brightcove, and also importantly YouTube. For example, the Cowboys could harvest select Cowboys-related YouTube videos and incorporate them into their site (this is similar to what Magnify.net also enables). With the chromeless approach, the Cowboys's user experience and their video player's branding is maintained while YouTube's rules, such as no pre-roll ads are also enforced.

To the extent that chromeless APIs become more widely available, it means that syndication can really flourish. The underlying content provider's model is protected while simultaneously enabling widespread distribution. All of this obviously leads to more monetization opportunities through highly targeted ads.

Bottom line: EveryZing's new MetaPlayer addresses at least three real hot buttons of the broadband video landscape: improved navigation, enhanced engagement and expanding content selection/monetization. All of this should give MetaPlayer strong appeal in the market.

What do you think? Post a comment now!

A couple of days ago, Truveo, the big video search engine owned by AOL, released the results of an internal study which concluded that it provides the most comprehensive search results among 5 companies considered. Before you say, "Duh, Will, what else would you have expected Truveo to conclude?!" it's worth spending a few minutes considering the study's methodology, results and implications. Video search is an extremely strategic space, so all credible data has value.

When it comes to search, there are really two key criteria to judge quality - coverage and relevancy. A search engine can return a million results, but if none are relevant, it's pointless. Conversely, just one spot-on result and you'll rejoice, but you still may yearn for additional, relevant options (since video quality can vary, links may be broken, the user experience at certain sites may stink, etc.). So optimizing both coverage and relevancy must be the goal.

In Truveo's study, it has focused solely on coverage, having deemed relevancy too subjective to credibly measure. To quantify coverage from a competitive standpoint, it chose 4 other search engines, Blinkx, Microsoft Live Video Search, Google Video and Yahoo Video. This limited pool immediately begs the question how the many other video search companies not included would have fared. Truveo explained that the testing was very resource-intensive, so they needed to keep the competitive set relatively small.

To measure coverage, Truveo selected 100 top-ranked Alexa sites across 5 categories: news, sports, TV, music and movies. Then they found 10 representative videos from each and ran a query for those videos - using the exact title the site used - on each of the 5 search engines. Scoring was binary - a search engine got a 1 if they returned an accurate result for at least 5 of the 10 queries, a zero if they didn't. Final score from this process, Truveo 86, Blinkx 20, Microsoft Live Video Search 17, Google Video 3, and Yahoo 2.

Having reviewed the test's full methodology and spoken to a Truveo representative, I think for the most part their approach is pretty fair. An obvious limitation is that lots of video search engines (or web search engines like Google) weren't evaluated so the study is by no means conclusive. Further, only premium sites were included (i.e. no UGC, and actually very little indie video either), so one wonders how the results would have changed if sites like Break.com, Heavy and others were also tested. And then there's the small matter of YouTube, the market's 800 pound gorilla, not being included at all. Since for many users video search begins and ends with YouTube, its omission raises a question about just how reflective these results are of real-world user behavior.

Nonetheless, Truveo gets points in my book for shedding further light on a very confusing subject, and also constructing a relatively objective methodology that can be used by others (in fact Truveo is encouraging independent 3rd parties to undertake more testing of this kind).

Video search is one of the most intellectually challenging areas of the broadband video ecosystem, yet as Truveo asserts, there is surprisingly little evaluative data out there. From my standpoint, more data means more informed market participants and therefore continually improving user experiences. That benefits everyone in the broadband ecosystem.

What do you think? Post a comment now.

(Note, the complete methodology can be requested by emailing Josh Weinberg at jweinbergATtruveo.com)

Another update from the chaotic video search and navigation space; Mefeedia, which has been flown below the radar, has quietly grown to 5 million unique visitors per month since its re-launch in January, 2007. I spoke to CEO/founder Frank Sinton yesterday to learn more.

Frank explained that Mefeedia hasn't focused on having the largest video index (its index contains around 12 million videos, compared to blinkx's and Truveo's 120 million+). Instead it has emphasized higher-quality, ongoing "shows," which it makes available as feeds to its users - hence the name Mefeedia. The company search from 15,000+ video sources.

Frank believes that the company's secret sauce to maintaining quality is its proprietary video crawler, which is modeled on Google's PageRank system. Mefeedia's algorithm looks at social ranking factors including embedding patterns, links and usage to build the index. This contrasts with others' approaches which look inside the video itself and/or search metadata.

Frank thinks that with 50K-100K new videos uploaded each day, Mefeedia's approach is more scalable. My take is that Mefeedia is cleverly playing on the 80-20 rule - far more users will be interested in premium, well-known or well-organized content, rather than random YouTube clips. In fact, Frank said that fully a third of Mefeedia searches are "navigational," i.e. using specific terms like "Family Guy" rather than general terms like "Golf."

Meanwhile, another clever approach Mefeedia uses to maintain the quality bar and also engage users is enabling them to curate "channels" of interest within the site. This involves a user/enthusiast sifting through various feeds to assemble one master feed, which other users can subscribe to. There are hundreds of these channels, naturally ranging from the expected to the very Long Tail.

Like all search and navigation sites Mefeedia is free and ad-supported. Frank sees four types of ad implementations: keyword ads around search results (currently from AdSense), related sponsors for specific channels (e.g. Kraft for cooking channels), banners on the site, and pre-rolls. In this last category, Frank said that the company is starting to do biz dev deals with content providers whose video would get additional prominence in exchange for Mefeedia gaining the right to sell certain ad inventory.

Mefeedia's progress since its re-launch from its roots as a Vlog directory is impressive, especially considering it's a 5 person shop which has raised only $250K in angel funding. Having proved its appeal to users, Mefeedia's next challenge is to prove it can monetize its traffic.

What do you think? Post a comment.

Metacafe, the short-form video aggregator with 30 million monthly visitors, has unveiled a new feature called "Wikicafe" which addresses the daunting and ongoing problem of how to find exactly the video you're looking for and gain high-quality recommendations.

Now in beta and available to its registered users only, Wikicafe is philisophically similar to Wikipedia, which involves users in building the knowledge base around specific content. Similarly, Wikicafe's goal is to involve users in continually refining the metadata for specific videos. This in turn will yield improved search and discovery for subsequent users.

Wikicafe is an intriguing spin on video search which I have discussed a number of times. Last week I spoke to Eyal Hertzog, Metacafe's co-founder and now chief creative officer, who's leading the charge on Wikicafe. This was the first briefing Metacafe has given on the new Wikicafe feature.

Eyal notes that there are really two ways to tackle content navigation. One is through super-sophisticated algorithms and distributed hardware, an approach epitomized by Google. The other is community-based collaboration, an approach epitomized by Wikipedia. He is biased toward the latter because he believes that the likelihood that the original metadata assigned by the video's creator (and even subsequent metadata that may be produced by technology-based approaches) will never be as accurate as that which is produced by other humans with specific domain knowledge.

Thus the idea behind Wikicafe: if given the right tools, Metacafe's users will create and maintain the most accurate metadata for Metacafe's vast collection of videos. It's a classic "wisdom of crowds" approach. Of course, it also requires that users act appropriately or things could spin out of control very quickly.

Wikicafe is very straightforward to use. Once logged in, you simply click on "Editing Options" in the upper right corner of each video. Then you can start editing the video's title, tags, description and then save your changes. You can track your changes (and those that others add), be notified about subsequent changes and start a discussion about your changes. You can even translate your changes into other languages. As Eyal explains it, this "collaborative taxonomy" allows redirection between related terms ("PS3" and "Playstation3"), clarifies ambiguous words, resolves hierarchical terms and connects different languages.

In a sense, Wikicafe is a natural evolution for Metacafe, which has always emphasized community involvement in filtering which content gets added and promoted on the site. With a group of active, passionate users and Wikipedia as a model, it seems likely that Wikicafe will gain traction in the community.

What then becomes especially intriguing is the potential for carrying the Wikicafe approach outside of Metacafe's borders for the larger universe of broadband video. Could users eventually become an augment or even replacement to top-down driven video guides, the norm in today's cable and satellite offerings? It's an interesting vision to contemplate. First let's see how Wikicafe evolves in the Metacafe community.

What do you think? Post a comment now!

As the proliferation of broadband video continues apace, the task of finding what you're looking for is only intensifying. That's set off a scramble by many to solve this problem, trying to become in effect, the "Google of video search."

I've previously written about players such as blinkx, Truveo, ClipBlast, Veveo and EveryZing. The latest video search company to hit my radar is Pixsy, which has its own distinctive approach for capturing its share of the growing video search market. I recently spoke with Chase Norlin, Pixsy's CEO to learn more.

Pixsy has three key differentiators:

First, it's purely focused on the B2B white-label opportunity, eschewing the destination site route. Pixsy only wants to power other sites' search capabilities as a white-label provider.

Second, in addition to video search, Pixsy also does image search, which Chase believes is actually the fastest growing part of the search market. Being able to offer image search broadens Pixsy's value proposition to partners, driving enhanced monetization opportunities.

Third, Pixsy doesn't seek to have the broadest index, rather it seeks to balance the breadth of its index with having the most current results. It uses RSS feeds, not web crawlers, to build its index and focuses mainly on current events categories like news, sports and entertainment. Clearly having the most up-to-date results in these kinds of categories is a real plus.

Pixsy's approach seems to be paying off, as it is now powering video and/or image search at sites including Veoh, Lycos, PureVideo, National Lampoon and others. This morning it's also announcing a deal with MMORPRG.com, the web's largest massive multiplayer online role playing site.

Pixsy works with a number of business models. Sites generating anything under 10K search queries per month can freely use Pixsy's API. Above that volume, Pixsy licenses its index with about a third of partners selling their own ads, and the other two-thirds relying on Pixsy to sell the ads alongside the search results. As Google and others have shown, search results are extremely monetizable.

Based in Seattle and San Francisco, much of Pixsy's team comes from Microsoft and ValueClick. Considering it has been around for about 2 years, has only 15 employees and has raised just an angel round, Pixsy seems to show that the barriers to entry for savvy video search startups can be relatively low. With so many other video search players, I anticipate the category is going to remain fragmented and chaotic for some time to come.

blinkx, which has been steadily expanding its portfolio beyond its core video search product, this week announced Advanced Media Platform or "AMP" (not to be confused with Adobe Media Player/AMP or Yahoo's Advertising Management Platform/AMP). I spoke with Suranga Chandratillake, blinkx's founder/CEO last week to learn more.

blinkx is addressing a problem that I hear about often - how can content providers which have increasingly large volumes of video on their sites make it more discoverable, helping drive usage and therefore ad revenues. Just this week on a panel at Digital Hollywood, Andy Forssell, Hulu's SVP, Content and Distribution, highlighted this problem, saying "we believe great content is significantly underwatched."

With AMP, blinkx has packaged up various offerings previously available web-wide into an enterprise product. These include its core video search/indexing technology, plus an SEO module and AdHoc, its contextual advertising platform for targeted monetization. blinkx is positioning AMP as a comprehensive approach that content providers can implement quickly on their sites. AMP is available in both licensing and ASP models. In the ASP model, AMP is available for a fee, or through ad revenue sharing. Suranga believes the ad sharing approach will likely end up being most popular.

blinkx announced 3 new AMP customers, Conde Nast's Portfolio.com, WallSt.net and Kiplinger.com. blinkx's AMP reminds me most of EveryZing, which I wrote about here. EveryZing's ezSearch and ezSEO take a similar approach to wringing value out of video assets. Pixsy is another company offering white label video search. Earlier this week it announced National Lampoon's network of sites as a new customer.

As content providers shift their focus from just getting their video online to actually monetizing and earning an ROI on it, discovery becomes critical. Therefore, I expect lots more activity in this space yet to come.

ClipBlast, an up-and-comer in video search, has quietly introduced its "Playbox" feature, which neatly wraps together video search and video syndication, two key mechanisms for increasing video usage. Gary Baker, ClipBlast's CEO gave me a sneak peek a few weeks ago and a demo earlier this week.

Playbox's key differentiator is that it provides a unified video experience across different video providers, while keeping users on-site. When clicking on any result which has video from the web that's embeddable or video from providers that ClipBlast has deals with, a floating window opens on top of the web page.

Playbox plays multiple video format (Flash, WMP, QuickTime) and correctly re-sizes each for the Playbox window. When you click back on the site and select another video it seamlessly loads right into the Playbox window. All of this contributes to a highly unified experience.

While that's pretty cool, the more interesting aspect of Playbox is that it drives frictionless video syndication. Here's how it works:

As ClipBlast distributes its video index and search bar to other sites, Playbox comes along (or not if the site declines it). This means that as ClipBlast's video search migrates from its site to the far corners of the web, Playbox's video syndication spreads too.

Here's an example: say you run a popular blog about auto racing and choose to offer ClipBlast video search and Playbox. A user of the site types "Danica Patrick" into the ClipBlast search bar (for those who missed it, she was the first woman to win an Indy car race, last weekend in Japan).

If the user clicks on the first result with a "Play" icon (with a rollover that says "play this video directly"), Playbox opens. In this example, the first video is from ESPN. When it's over, the user can click the "Details" button and a list of additional, related videos are displayed, from other providers such as AP, VidLife, YouTube, YES Network, etc. The point is that Playbox provides an easy-to-use container where related 3rd party videos can be played without the user ever leaving the auto racing blog.

So Playbox gives videos new exposure throughout the web. Again, if providers code their ads into the file, more views = more revenue. The Playbox window also shows small display ads at the bottom, which ClipBlast monetizes through ad networks. Depending on the video and search term, these slots lend themselves to contextually-targeted ads.

Playbox marries video search and video syndication, and is another mechanism for how video is going to be widely distributed around the web. I expect more of these initiatives to come.