23/02 : Phorm on TechDirt
Yes, TechDirt have picked up this story and are running with it. Their coverage is fair and does provide an interesting link to the formal announcement by Virgin Media, BT & Talk Talk of the Phorm partnership. More notable however is the reply posted to the comments on the article, supposedly from a member of the Phorm technical team. If this is true, it's the first public response we've seen from Phorm to the almost universally negative press this announcement has been getting.
Read the TechDirt article here :
http://www.techdirt.com/articles/20080218/024203278.shtmlComment 22 (reproduced below as comments have a tendency to disappear over time) claims to be from Phorm's Tech Team :
Relevance and Privacy protection by Techteam at Phorm on Feb 22nd, 2008 @ 3:01am
Hello Mike, I’m a member of the Tech team at Phorm. You know - there’s no way the AOL situation would happen with our technology. The OIX throws away raw data instantaneously so there’s no click stream history to be inadvertently released. I know what we're doing seems out of the ordinary from what’s been seen previously in the online Ad space - so I just wanted to make a couple of points to clarify how the technology works. After all it is counterintuitive to claim that you can present consumers with more relevant ads without collecting and storing personal information - but that is exactly what our system does. If a subscriber agrees to participate – maybe because they are tired of receiving ads that have no relevance to the products and services they want to buy – they get a random number. So they are anonymous to the OIX. As they browse anonymously they match advertising channels that have been defined by patterns of URLs, keywords and search terms. But what’s so neat, is that even before each page loads the information observed from that page is deleted. The only data left in the system is just a random number and the channel it matched. In addition the OIX only looks at information relevant to the ad channels - so it doesn’t scan names, numbers, emails, secure pages and form. Nor does it allow ad Channels to be constructed for a range of sensitive areas like adult content, alcohol, medical - so again browsing behaviour in these areas simply isn’t observed. Since there’s no personal data stored – nothing can be traced to an individual and there’s no information to reverse engineer. There’s no interface or database in the system where a UID can be entered to retrieve profile information. We understand that even with all these safeguards - users may not want to participate, and that's alright. There's a simple option for users to opt out on webwise.com. We even tell them to how to block the domain for a more permanent option. The team here has really worked hard over many years to build the system from the ground up with privacy at its core. We’ve also worked closely with user privacy groups in the UK to make sure that we did this right, and that user's privacy is protected. I'd be happy to discuss any further questions you may have. Pop over a note and I'll send over my contact information. Cheers, Tech team 