Cesar Brea's Weblog My original blog, hosted by the Berkman Center for Internet and Society at Harvard Law School
Octavianspace A Myspace experiment. May 2006 update: no friends after 6 months (Tom doesn't count). Maybe this isn't for me, though I haven't done much with it yet.
Marketspace Advisor News and comment on the cross-channel customer experience
Radio Free Brea My podcast station on Andrew Grumet's Gigadial service.
ESM Partners essays on high-tech strategy, sales, and marketing by me and Jamie Schein.
A while back, I wrote a post suggesting that Google's AdSense business was less profitable than its direct search business. Some folks I know in the business have confirmed this. Yesterday, the New York Times carried an article about how Google is trying to keep people on Google.com longer, and making publishers and retailers mad in the process. But what can they do about it?
Dana Todd and Bruce Clay on SEO. Interesting to SEO nerds and newbies alike. The comments about getting "theme" right were especially interesting. At a deeper level, the ideas here are metaphorically extensible to marketing in general.
Given the preceding post describing the hype around ad networks, I thought a closer look at the uber-network might be useful. What I found surprised me.
Google reported $10.6 billion in revenue for 2006. Of this amount, $4.2 billion came from placing ads on non-Google-owned sites. They paid those sites' publishers $3.1 billion in "traffic acquisition costs", resulting in a gross margin from their AdSense ad network business of $1.1 billion.
In 2006, Google spent $1.2 billion on R&D, a little more than $800 million on sales and marketing, and a little less than $800 million on overhead, for total operating expenses of about $2.8 billion.
A simple allocation of this $2.8 billion to Google-as-media-firm versus Google-as-ad-network based on each's contribution to total revenue puts 40% of these operating expenses, or $1.1 billion, onto the latter. The resulting math makes Google's ad network business a break-even proposition in 2006.
A week ago I
had lunch with Waikit Lau, president and co-founder of ScanScout (http://scanscout.com). Waikit, whom I got to know when I was at ArsDigita and he was part of the http://photo.net team a few years ago, describes ScanScout's video analysis technology like this (this is the shareable version):
For the NHL, Gotuit transforms what used to be a 60 minute linear viewing experience to one that can be sliced and sequenced in a variety of ways, without having to cut, splice, or otherwise edit the underlying video asset.
The social shopping service Kaboodle recently announced the availability of its "Help Me Choose" Widget. Essentially, this allows Kaboodle users like me (my wife and I use it to maintain a shopping wish list for home and kids) to publish a poll on external properties like this blog, so friends can offer input into certain purchase decisions.
"Yahoo! Answers Adds Social Networking", writes Loren Baker at Search Engine Journal (nice demo video from the Yahoo! Answers team included). With everyone trying to build an online community to grow these days, what to make of Yahoo's latest gambit?
A couple of weeks ago I had breakfast with Patrick Donovan, and old colleague from when we worked together in the early 90's at a Lexington, MA consulting and software development firm called Symmetrix. These days Patrick is VP of Product Development at Gotuit Media. Gotuit provides technology that allows users to "deep-tag" a slice of a video they look at on sites like YouTube and Metacafe.
Gotuit doesn't actually rip and store video itself; rather, it's an interface through which you create a data layer (on Gotuit) that identifies and describes (with title, tags, and a free-form text field) slices of videos hosted elsewhere. Gotuit provides a browser toolbar with buttons that make it easy to quickly deep-tag a slice of a video and then share it via embedding or a hyperlink in a blog or other web page, or by simply emailing it to one or more friends.
So what? Search is already a killer app on the web. Video is exploding on the web. Ergo, search for video will be huge. Since video (and audio) is consumed linearly, meaning you can't browse it the way you can browse text, simply tagging and otherwise describing a source file on YouTube or somewhere else is only partially helpful to getting you to what you're looking for and helping you consume it efficiently. Being able to deep-tag slices becomes really useful particularly for form factors and contexts -- like mobile -- where efficient use of limited resources (time, bandwidth, and memory) are more important.
It doesn't stop there. Of course, if you can tag inside a video, and you can sell ads against tags, you can advertise inside a video, in a highly targeted way. And if you can tag inside a video, you can also string together slices to create what we used to call "highlight reels", but now would call "tag-dimensional mashups" I guess, allowing "omni-directional re-purposing" of content. Videos of NFL games could be sliced and mashed up to create "best touchdowns", "hardest hits", "ugliest players" series. Again, particularly interesting for mobile applications and contexts.
While I'm not sure that Gotuit in its current business model incarnation (today's separate service vs. a licensed, embedded capability in major video sites) or UI expression has reached its final stage, this is a really interesting company, and Patrick is a very smart guy. Check them out.
