10 conditions of success for online video

Today, I received my daily email commentary from my friend Matt Coppet, who runs global media and strategy at UBS. He & I occasionally discuss the new media space & share ideas, and I wholeheartedly agree with his thoughts today. His commentary is reproduced below.

Since we published our "10 conditions of success for online video" three weeks ago, new announcements came in. CBS started the CBS Interactive Audience Network and syndicated its content to ten online channel partners. The NBCU-NWS partnership signed up Comcast and added new cable networks to its future line up as well as offering new distribution outlets through Comcast online sites.

CBS and the NBCU-NWS venture offer diverging economics but aim at a similar outcome, the creation of a new syndicated marketplace, matching growing online content inventories to currently strong advertiser demand.

*CBS decided to give up some of its distribution control (and surrounding environment) for broader penetration (working with all the potential core players such as Joost, Brightcove, AOL, Microsoft and Veoh). By broadening its online distribution sooner (using IBM "hyper-syndication" term), CBS could forgo some potential monetization in the longer term and suffer conflict with its brand (some of its partners have varying traffic quality).

*The NBCU-NWS venture has a more daunting task, which is to generate user destination traffic. The capability for the venture to invest in marketing the new site(s) will be key. Comcast's ownership of several high profile technology partners should facilitate that work. Our conversation with The Platform (Comcast owned) CEO Ian Blaine highlighted many potential synergies, mainly for content management and syndication support. Our conversation happened before yesterday Comcast announcement.

Implication => We believe a syndication model would enable studios to regain eye balls online. It is too soon to estimate how user generated content (UGC) demand will evolve following the massive arrival of easy to access, indexed and searchable studio content (both in long and short format). AOL, NBCU-NWS venture, Joost (expanded beta in May) and other platform would likely take market share away from current UGC leaders.

We highlight again our 10 conditions of success for any online video ventures: Traffic optimization, advertising integration, brand focused, syndicated model, social media friendly, independent decision making, bandwidth costs control, leveraging data mining, bit sized content available and predictive marketing

(1) High web traffic from start is necessary to sustain advertisers' interest and to soften the impact of higher ad inventories on CPMs. Current CPMs up to $50 would be tough to retain in a basic streaming environment (no targeted ad).

(2) Integrated ad servicing process is key (no ability for each partner to sell its own ad inventory, which would defeat the purpose). It also eliminates channel conflicts.

(3) Brands/advertisers friendly (brands should be able to go behind pre roll/post-roll ads. We believe opening the full interface to branding/ sponsorship - including the player skin - is key. Ability for users to bookmark content and distribute segment is important.

(4) Syndication model can work with compelling content. It is not a surprise to see NBC Universal in the venture. It recently pre-launches the National Broadband Co. (nbbc), a content syndication marketplace. We do not know yet if nbbc would be integrated into the venture or become an infrastructure partner.

(5) Minimal social-media (web 2.0) function (such as users' commentaries/ ratings/chat). Over promising in term of user customization could conflict with future branding/sponsorship. For instance, mySpace users would be able to embed those videos from the NBCU-NWS venture on their webpage. Any venture would have to make sure of the integrity of the player branding (against program scripts that could remove them).

(6) Independence is important. Operating power should not be shared among partners, which could slow down the decision making process. It also eliminates conflict of interest.

(7) Contextual advertising (or predictive marketing) should enable advertisers to focus on branding as well (and not only on shorter term promotion). As relevance improves, brands would likely spend more driving monetization up.

(8) Managing bandwidth costs is central to any ventures success.

(9) Moving from simple data gathering to leveraging data. Understanding the customer (analyzing databases) offers a potential viewer "lifespan value". Package goods companies for instance could be inclined to spend more for addressable advertising.

(10) Offering content in the format customers want (such the 5 minute signature moment in a sitcom or a specific candidate performance on American Idol). Smaller length would respond their younger demo aspiration and could increase the ratio ad over content.

How to Marry a Millionaire

Now that in-game advertising is gaining traction, the next steps would be to allow marketers and advertisers to target their audience with the most relevant messages. How would they do that? Predictive Analytics to the rescue!

As per Wikipedia, Predictive Analytics are techniques that use past performance to assess how likely a customer would display a specific behavior in the future. An example application is credit scoring in financial services, where attributes such as credit history, loan applications, payment defaults, etc are used to assess how likely an individual would make future credit payments on time.

Now that the discipline has matured in one industry, it's time to port that to the new media world, particular in-game marketing and television guidance. A Business Week article suggested that predictive analytics would provide the intelligence to help marketers target consumers while playing online games. By analyzing historical campaign performance, and hundreds of gamer attributes, including session activities, downloads, number of friends and genre preferences, marketers can identify consumer preferences and behavior. In an online gaming, different messages can be targeted to different players, each getting a message relevant to them. With such targeting, marketers would be willing to pay a higher price (per capita) to reach the right consumers. Moreover, since social networking enables peer-to-peer influence, predictive analytics can also identify the influencers who have loyal fan followings, not unlike identifying the most popular kid in high school, only through data analytics.

In-game advertising is only one application. That discipline can be used in television guidance. Then additional (though more targeted and more relevant) advertising opportunities can be created. With personalization and recommendation engines powering TV Guide 2.0, why not employ predictive analytics to (1) add to the power of giving the consumers exactly the television (or video, in general) that they want, and (2) market to them at the same time, since we understand their viewing behavior already?

For The Love of the Game

In a recent eMarketer research report, Video Game Advertising, advertising spending in video games (either by in-game advertising or advergaming) is expected to grow by 23% CAGR (compound annual growth rates) annually in the next 5 years. In the US alone, such spending is expected to almost triple between 2006 ($346MM) to 2011 ($969MM). Advertisers are finally realizing where their target consumers are and going after them there. With such tremendous growth in ad spending, it also means that (1) more consumers are playing video games than before, and (2) that more than young males (ie women and older gamers) are being targeted.



More players mean fewer of them are watching television, the advertisers' previous playground. Looks like it is no longer enough for a television network to offer broadband videos anymore, they will have to include a gaming strategy as well, if they want to keep their piece of the total advertising pie. Google bought Adscape recently for $23M and bought Massive for $200MM, both of which insert advertising into gaming environments. It's only a matter of time before they make it easy for advertisers to buy into video games and with better analytics.

While console games still skews young, adults prefer casual online games. According to WSJ, Forrester Research reported that over 50% of online games players are over 30 years old. Their favorite games include mahjong, Sudoku and solitaire, and the gender breakdown is evenly split between male and female players.

Even print publications are getting in the "game". In the same WSJ article, teen titles from Hearst Magazines such as CosmoGIRL! and Seventeen had been getting up to 10% of their site traffic from casual online gamers. Understanding that fact, they are redesigning the site to feature casual games more prominently to increase and engage their audience base. Other magazine websites are following suit. As casual games become mainstream and are easily available, their players will have to trade out other entertainment time (ie television, movies, etc..) for games.

Where's Waldo

Recommendation or search? The battle is just beginning to help viewers discover their video content.

Recently, two online personalization and content recommendation companies received 10 figures funding each (ChoiceStream raised $26MM for their 3rd round; Aggregate Knowledge raised $20MM in their 2nd round). While at first glance, these P&R companies many use their new funds to sell their services as online behavior targeting uses (AK specifically earmarked their $$ for sales and business development), it would only be naturally to extend these uses to online video.

Why online video? Because it can be a trojan horse to the home, particularly if the personalization and recommendation engines can be ported seamlessly across multiple platforms (broadcast via set top box, online and mobile). The software that powers these platforms can engage the viewers, which then leads to better ways to advertise, market and sell to those viewers.

TV networks are pushing a lot of video content to their sites and at NATPE this past January, many television executives verbalized their valid concerns for video discovery. As more video becomes available online (especially from television networks), online viewers will have a tough time finding videos that interest them. While video search engines (Blinkx, Truveo, Google Video, etc..) are now coming of age, they still dependent on good metadata and tagging taxonmy, both of which are managed by "gatekeepers", to enable video searching by keywords. Keyword searching may be perfect for text content, but would it work for videos as well?

iTunes, Netflix and Tivo recommend video titles based on past experiences (ie ratings) by oneself and/or from other similar raters (ie friends). In these cases, searched content is specifically restricted to within each of their walls, which then returns the most relevant recommendations without the long-tailed outliers corrupting the results. Taking such cues, TVGuide is launching phase 1 of Project Stingray that limits its search to only 65 relevant video sites. I can only postulate that phase 2 would incorporate personalization and recommendations to video searches, with more relevant results given heavier weightings. Viewers searching for "The Office" are more likely to look for the television show rather than homemade videos of your co-workers' shenanigans at the office. However, by drawing a distinction between the two, amateur gems can still be discovered and the integrity of user-control viewing is maintained. It is “TVGuide” after all.