
I just came across this interesting new testing approach on the YouTube home page that allows website visitors to opt into a test and also explains what's being tested instead of just randomly allocating people into test groups using the standard a/b/n approach.
Giving choice to people is usually a good thing but in a testing situation that's kind of new and I'm not sure how I feel about this, what are your thoughts? Using an opt-in approach and explaining the test should generate more buy-in and maybe even unprompted feedback but I wonder what the non-random allocation of test subjects does to the significance of the test results.
This is a must watch video for all data fans, it includes a few cool data visualisation examples. David also talks about data being the new oil and tries to coin the new phrase of "data being the new soil" (i.e. insights and innovation spring from it), great idea I think but it comes across as a bit wanky. Anyway, awesome examples and great video!
Amazing source of information to get a deeper understanding of what triggers your audience interest. Below is a sample output showing me videos that 25-34 year old males in Australia like to watch including search terms used and how well they index by category. Nice!
Try the audience insights tool here
http://video-analytics.google.com/yap/iba
Or check out the other beta options
http://www.youtube.com/testtube