Showing posts with label data-mining. Show all posts
Showing posts with label data-mining. Show all posts

Thursday, May 03, 2007

Google History

Have you checked out Google history? You should. It is awesome, and brilliant and scary, all at the same time. If you have read my blog before and wondered about my posts on Attention, or went and read some Goldhaber stuff, or checked out Atten.tv, and you still find that this topic is a bit academic or difficult to grok, then I suggest you go check out Google History.

A week or two ago, I read a post about Google history and quickly signed up and allowed the logging of all my data in a variety of categories. Then I forgot about it, much the same way I keep my Attention recorder running on my browser (By the way...When is Root.net going to fix their security certificate? Very disappointing, but that is another post...).

Earlier this week I went in to see what was in my Google history goodie bag. The main page is a chronological listing of all the pages I clicked on from now going backwards. I have this in Root.net as well but I get to block out certain domains I may not want to record (things like ego surfing my site statistics). Not here, everything gets recorded. Not really a problem, but still kind of freaky when you sit down and think about it and then multiply me by the population of the web world who uses Google. I am creeping you out yet?

I can then slice it more finely by looking at just news, or images or videos etc. especially for Google related services. Not too shabby.

I then click on trends which slices out the top queries, top sites, and top clicks. Are you still wondering why they are making so much money in advertising? Hold on there is more...I almost forgot to add you can slice that by week, month, all time etc.

My favorite part though, and where I think it all really comes together is the tab called interesting items. Here Google has determined other searches, pages, videos and gadgets that might be relevant to what I have been up to while surfing on the web. The list of pages is pretty impressive as we have been doing the fund raising exercise and the top 10 list for me includes nine venture firms and a buyout firm. Uncanny. I am curious though on how they rank them. Could it mean that the sum total of my interests (wireless, media, etc) when intersected with venture fund raising means that these are the best prospects for closing a round with those specific venture firms? Hard to say. In any event I was pretty surprised that there weren't fantasy football references or NFL draft results or other items which given the recent NFL draft would make the most sense. Maybe the algorithm is skewed?

In any case, if you have the slightest interest in Attention or wanting to understand the data you are throwing off when you mindlessly browse and search the web, I would certainly recommend that you check this service out. In any event, Google isn't the only company tracking this data (Yahoo, Doubleclick, AOL, etc.) and I think it is important we all be cognizant of the invisible transactions we are all generating behind the scenes with our web based actions. A final note is that I am not a Google hater or fearer as I know some are. I use Blogger, Google, Gmail, Google Apps and occasionally Google Maps. I get that the price of the services I am using are a value exchange that they need to offset by profiling me in great deal to yield high advertising returns. I would however like for Google to give me the option to keep my own copy of all the data they are keeping on me in the event that I want to use it for my own benefit somewhere down the road. You know, like when I am trying to decide which venture firms I should call on first, or which ones I should cross off the list. :-)

Monday, April 09, 2007

Atten.tv

While away at CTIA a couple of weeks ago, the annual eTech conference took place in San Diego, which I of course missed. One of the people unveiling a product there was Seth Goldstein who announced Atten.tv. This site and software builds on some of the earlier attention work such as the badge on this blog from Root.net, and now has an alpha software client for the Mac that is sort of like a television screen that shows you the various pages that someone is viewing or has viewed while recording their attention with tools like the Attention recorder. The alpha is very alpha in my book and I am anxiously awaiting the full beta.

As mentioned in my previous post, I cleaned up the various widgets on my blog and switched to a new template. Over the weekend I realize that I forgot to include urchin analytical tracking and the root.net badge that I have added now. I don't visit urchin as often as sitemeter but it is definitely another tool that I spend time with when trying to understand my audience or accidental visitors.

If blogging is a fame based or attention seeking medium, then one of the critical pieces to measure the effectiveness is the inclusion of analytical tools to document what is happening as well as tools to share that data with others. I personally like the root badge as it is a way to share at a high level, your activity with others, to the degree you feel comfortable. The development of the viewer (called the Attentron), is a vivid way to demonstrate that aggregate data in a more meaningful way. I have no doubt that this is yet another step in the development of rich tools that will form around the concept of attention going forward.

Friday, March 02, 2007

MP3.com Data Mining Revisted

Over two years ago I wrote a post about some of the data mining we did at MP3.com and then Vivendi Universal that you can find here. To this day I find that entire exercise fascinating and one of the things that I enjoy the most is that the story only gets better with time.

Back in that post I referenced Rick Walker who worked with me trying to identify opportunities to sell data to the record labels back in 2003-2004. The previous post talks about some of the mechanical specifics but it doesn't really do a good job showing what he had developed.

Around the end of the year last year we needed to bring in a new finance person for our current company and coincidentally Rick became available. It's been great having him on the team and we like to talk about how cool that project was and how far ahead of it's time the data was. This is especially poignant as we have seen bands that were essentially baby bands at that time that have blow up in a big way. On the other hand there are also a bunch that became roadkill. This morning I uploaded a sample of the report that we were sharing with A&R folks before Vivendi divested the assets of the various music properties I oversaw. Here is a link to the National chart and some of the larger markets. You can find three other pages here, here, and here. Take a special note of the top band in Chicago in 2003.

Thanks to Rick for dropping this blast from the past on my desk this morning. I have been chewing on a post for a couple of weeks ever since I had a pint with a BBC producer in New York who spoke with me at length on the challenges they face in the future in terms of getting paid. I don't know that I have any answers to that question but I think the framework of my thinking to describe the key drivers and point to some potential ideas is almost complete. As there are a lot greater minds than mine thinking about that problem I don't expect to break any new ground but I think this past experience when colored with some of my current experiences will at least be worth sharing or start some discussions.