Google P2P Network? (or, Its Easy to Score Relevancy When You OWN the Network)

Google, already has a near infinite number of data points to compute relevancy for the active parts of the web, and is looking to gather even more user data information. The WSJ has background on the story:

Google is preparing a service that would let users store on its computers essentially all of the files they might keep on their personal-computer hard drives -- such as word-processing documents, digital music, video clips and images, say people familiar with the matter. The service could let users access their files via the Internet from different computers and mobile devices when they sign on with a password, and share them online with friends.

They also mentioned the C word:

Google will likely have to address copyright issues. Allowing consumers to share different types of files such as music with other users could trigger the sort of copyright complaints the company already faces over videos on its YouTube video sharing site. One person familiar with the matter says Google is discussing with copyright holders how to approach the issue and has some preliminary solutions.

This is going to move Google up the value system by

  • giving them a unique data source
  • giving them unique relevancy signals
  • keeping users locked into their services and using their services longer
  • shift power from copyright holders to Google
  • eventually allow Google to sell content (if they want to - the Google Video trial did not work too well)

But there will also be a big upside, especially to marketers and content creators who are willing to give away high value content to gain mindshare and marketshare. By creating content that people would want to store and share on Google, you get cheap or free exposure for your business interests.

As DaveN said, Google eventually has to move away from links because links are too polluted. What better relevancy signals can they come up with than attention data and how often people cite and share data ON THEIR NETWORK? Feedburner, Google Reader, iGoogle, Gmail, and Youtube are already part of the Google network. Soon your hard drive will be too.

Published: November 27, 2007 by Aaron Wall in google


November 27, 2007 - 9:06am

Hmmm a Google P2P, that alone kind of gives me the shivers. I don't really know why Google would want to do this other than to gain even more data about people. If they end up controlling too much copyright power how long is it before we see the Google record label or G-Studios?

While this could be a potentially positive thing for young filmmakers or budding entertainment industry players I don't see how it would help the "common user" or indeed the poor writers already on strike in Hollywood.

There is a hell of a lot to be said about copyright/piracy and I don't think we really need to go into that on here but I do think Google would find a way to get away with it.

The comment about Google judging a site on network relevancy is very interesting but I don't think this is as viable as links because its much more easily manipulated, in fact in the case of GMail it almost encourages SPAM. I don't think Google would want that.
As mentioned however they do need an alternative, exactly what this is going to turn out to be is anyones guess.

November 27, 2007 - 9:49am

I don't think this is as viable as links because its much more easily manipulated, in fact in the case of GMail it almost encourages SPAM.

I have to believe that Google has some way to normalize account histories. Ultimately they can profile the business traveler, the entrepreneur, and the fake account based on account footprints. Everything from IP tracking, to iGoogle usage, to search history, to deleting spam in emails, to sending emails to trusted Gmail users, to getting email back from other trusted Gmail users...given that pool of data 99%+ of fake accounts would be easy to detect and filter out, IMHO.

You or I may not agree with how Google defines spam, but when they dig deep into a particular area they can find relationships that few others could tie together. They are pretty damn smart and efficient...just look at their profit margins.

Also, if you are uploading spam to Google, you have to create a Google account to do so, which is yet one additional footprint that will make it easy for Google to footprint and wipe out large swaths of spam.

November 27, 2007 - 1:34pm

It's true that Google are very efficient at removing their perceptions of spam, I don't for a second mean that more than 10% the spam I mentioned would hit users, i'm simply suggestion it would encourage people to try and manipulate things in this way and put a lot more work for Google and give people an excuse to generate more spam across their network.

Of course their is an argument that the current system invites spam (directories and link farms) so really I suppose it would be shifting the spam slightly closer to Googles control, even if there was more of it to be filtered out.

I really like my iGoogle and GMail accounts but I am very conscious of the data Google collect from it and i'm not sure how much I would want to use them if I knew it was having a direct effect on SERPs (of course for all we know it already does.) As you mention Google are very very good at tracking and analysisng data so whats to stop them noticing;

1)I follow you RSS to this blog a few times a week

2)Post on a few Webmaster forum and visit my site a hell of a lot

3) Say "This guy's an SEO, that means his sites are spam" and dropping me from the SERPS

Keep in mind Aaron i'm not a Google conspiracy theorist, I kind of like the Big G for all the positive work they have done in the SE field, i'm just not comfortable with them using their subjective opinions to control everything to the nth degree.

November 27, 2007 - 2:34pm

Looks like you included the same quote twice. The second quote is just the first one minus the first sentence. (Feel free to delete this comment once you've seen it.)

November 27, 2007 - 3:07pm

Thanks for the heads up Shane. I fixed it.

November 27, 2007 - 4:49pm

Google is going to become a statistical giant!

November 28, 2007 - 12:00am

The usage data thing and algorithms possibly shifting away from linkage data is one of the things I find most interesting in SEO. I know what you wrote about it in your book, but after that I think you also stated on here that user data would be able to be manipulated rather easily, too (not sure if I remember this correctly?).

Do you think if algorithms shift away from links to usage data that links might become rather unimportant? or would usage data "only" be an additional (even if very strong) factor in addition to links?

I mean of course links have become dirty, but wouldnt the same thing happen to usage data and people trying to manipulate that just the same way after a while?

I could also see them do stuff as they might be doing with the attention-whoring link bait thing and devaluing links that are not true "votes". For example I could see them trying to get better at analyzing the words surrounding (or in) the link (so that they won't count links that say: "look at this ***** ****. I absolutely hate this guy (yet I dropped him a link b/c he tricked me into doing it!)"

One thing I'm really wondering about is if web analytics skills will become a more important part of SEO once usage data becomes valuable. I mean web analysts use metrics to measure engagement of visitors with a site, etc.. And I guess Google or any search engine might be using similar metrics to make sense of the user data they accumulated (which might mean that it'd make a whole lot of sense for an SEO to understand what exactly they do). What do you think of that theory?

November 28, 2007 - 7:48am

Hi Patrick
I think they don't have to entirely drop links or entirely shift to usage data...they can just mix another set of signals into the algorithm.

User engagement will be a big thing going forward from a search relevancy perspective AND from a marketing / conversion / ROI perspective.

Add new comment

(If you're a human, don't change the following field)
Your first name.
(If you're a human, don't change the following field)
Your first name.
(If you're a human, don't change the following field)
Your first name.