October 30, 2007
Compete.com have released some stats listing the top 50 website domains by unique visitors on their blog today. It makes for some interesting reading:
Yahoo is still the biggest domain in terms of unique users. Not surprising given their huge coverage, surely they have to come up with a way to make a success of all these eyeballs? They may lose out in search to Google but with such a vast web real estate finding a way to leverage that is key for them. Google however coming second is amazing considering their core is still search!
Facebook at number 21 is a bit of a surprise, I’d assumed they’d be higher given the buzz but perhaps they’ll position much higher next year (if their bubble doesn’t burst).
The growth figures in the blog post are most intersting, showing sites such as YouTube, Flickr and Digg as some of the biggest gainers (bigger even than Facebook). This certainly is the time of sharing content, something Facebook has yet to get right (they started off well but it’s lately disolved into MySpace’esque profile vanity).
Adult dating still a major growth area it would seem; the person who launches a Facebook for this domain will win big!
Of the losers, most intersting for me is the losses experienced by Expedia. This can only be down to the emergence of much better sites that give users more intuitive ways to search for flight & hotel availability. Online travel is much more competitive in that arena this year and with the move from tour operators to embrace dynamic packaging I can only see Expedia losing more eyeballs if they don’t make some significant functionality changes soon.
October 16, 2007
At long last Google has announced the launch of some technology on YouTube which aims to protect copyright holders. Called YouTube Video Identification, the system does what Google has been promising to do since June, giving content owners the power to block copyrighted clips as they’re uploaded to the popular video-sharing service.
Of course, Google’s post on their blog avoids the mention of ‘blocking’ content and focuses instead on how publishers can ‘manage’ their content ‘proactively’. They also mention all the other steps they have taken to allow content to be managed on YouTube:
- Our strict repeat-infringer policy, which has been in place since our launch, terminates accounts of repeat infringers based on DMCA notices.
- We take a unique “hash” of every video removed for copyright infringement and block re-upload of that exact video file prospectively.
- We require a 10-minute limit on the length of content uploaded to the site.
- We provide content owners with an electronic notification and takedown tool, to help them more easily identify their material and notify us to take it down with the click of a mouse.
- We also publish copyright tips for users in plain English and clear, prominent messaging at the time of user upload.
Great propaganda guys… Here’s the details from the YouTube site. Will this be enough to fend off the copyright subpoenas? I think it probably will. This kind of proactive filtering approach is exactly what was needed to secure YouTube from the lawyers. Hopefully Google will continue to evolve this technology and stay ahead of the game with it, so securing YouTube’s future and allowing them to focus on the more important issues of monetising the video sharing site.
October 11, 2007
Google has just announced a new layer for Google Earth which features geotagged videos from YouTube. It’s great! As a frequent traveler and someone who works in the travel industry this is just what I need to discover new destinations.
I’m hoping that Google open up an API to YouTubes geotagging so we can integrate videos into our Google Maps implementation as well.
And while on that subject, what would be really useful would be if YouTube could add categories to it’s videos (such as travel), having access to those sorts of categories through API’s would make a lot of businesses much more willing to integrate YouTube content into their web offering (as you can avoid more of the dross).
September 18, 2007
Creating a visual search engine that can find video (or images) of people is a bit like the holy grail (highly desirable but tricky to get your hands on). Now along comes a company called Viewdle, who are one of the start-ups at the Techcrunch 40 event at the moment. Viewdle quote themselves as a ‘facial recognition powered digital media platform for easily indexing, searching & monetizing video assets’. Quite a powerful thing if they get it working right. Perhaps YouTube should buy them (seeing as Google have yet to make real use of Neven Vision technology).
Viewdle comes out of Reuters Labs and there’s a demo up and running which looks pretty cool. It shows how many times someone appears in clips and how long they feature (demo link here for Angelina Jolie). I’m hoping that as it comes from Reuters it should stack up commercially as I’m sure they’ll have used this on their own archives.
July 31, 2007
Some insights into the upcoming YouTube saving copyright filtering technology have been posted on the NY Times technology blog here. The text is a transcript from a session of the Viacom court case and is of one of Google’s lawyers trying to explain how the video signature fingerprinting technology will work.
It’s all as expected but makes for interesting reading.