You'd have a hard time actually reconstructing a life from a search history, because there would be so much noise. For every search that reflects an important life decision, there will be dozens (hundreds? thousands? I can't estimate very well) that you entered out of idle curiosity, random whims, etc. The significance of the ad is that those important searches will take place. Google will have no way of knowing which ones they are.
Yes and no. I have 18000 searches in my history, so I agree that you can't easily go through them all and trace out my life. And yet it is very simple to search it for trends, key events, or even pick a day and see what was going on then. I can do that, and Google can certainly do that (algorithmically of course). Now, I don't have a problem with that (and I value the ability to look back through my own life), but it does represent a trove of personal information.
Google's data-mining fu is extremely strong. This is what they do for a living. I'm sure they would be able to reconstruct a life if they choose. For every important life decision there's probably more than one search involved (having a baby? you'll search for crib, babyfood, baby crying, how to change diapers and much more) and it is certainly possible to pick that out from random searches for cribs because of boredom, checking it for a friend, etc.
If you move (change IP) Google will be able to track you if you use any of their sign-in services (example@gmail.com now regularly logs in from this IP instead of the old one). Your only chance of breaking this chain is either to regularly use proxies or to change IP and gmail (and other services) simultaneously.
This is basically the plot of Caprica. Zoe created a search engine so powerful that when she died, it was able to reconstruct her personality from publicly available information about her. Then her father downloaded her into a robot, and thus was born the first Cylon....
You'd have a hard time actually reconstructing a life from a search history, because there would be so much noise.
Did you see the analysis of the AOL data leak? There is one particular woman in there where consecutive queries and a little editing show a story of her relationship with an abusive boyfriend, accidental pregnancy, abortion, and remorse.
People found it by, essentially, just grepping a flat text data file for terms they knew were likely to resolve to a hot button issue, then separating the wheat from the chaff. This is the "solution you can knock out in vi in ten minutes" answer. If you have a team of Google-style engineers, a large volume of search queries with unique IDs on them, and a Hadoop cluster available, you can make sift-and-winnow as many gigabytes as you have down to as few salient details as you want.
Trivial example: I would stake money on being able to correctly identify the gender of over 80% of users given 100 searches from them, and I would be able to do this with less than one day of programming. If you let me do it in other than realtime, I would be able to do this at terabyte scales using EC2 and Hadoop. And I'm just a guy who did natural language processing in college and has a bit of basic web-fu available. Think of what you could do with a talented engineering team, deviousness, and a few months to work on it.
I just went through my last two months of Google Web History, and I can tell you, it doesn’t reveals anything that anyone that know me doesn’t know. The fact is that every day web searches are not as interesting as the ones in the ad.
Most of them in my case are software development-related stuff. Before being so afraid of “the power that google has” just check you history: it’s less frightening than what commercial made you believe.
Signal v Noise - suddenly having your server overwhelmed with 'interested superbowl watchers', and then losing people actually looking to buy your product is not so nice.
But (almost) any publicity is good publicity, and this may inspire a few people to fly to Paris / get married in Paris / make a DIY crib etc, which has positive flow on.
I don't own a tinfoil hat but that was pretty creepy. The idea of this corporation portraying itself as some sort of buddy who will be by my side as I go through life...
To me it comes across as a mixture of arrogance and over familiarity. Google wants to know us better so they can serve more targeted ads. That desire is not reciprocal. Not for me at least. Its like it never crossed their mind that people may not be as pleased as they are at the idea of an ad network being their best buddy.
> "The idea of this corporation portraying itself as some sort of buddy who will be by my side as I go through life..."
But often they are. Example: I grew up watching James Bond movies - I still watch James Bond movies, and there will probably be many more Bond movies between now and when I die. It's a known quantity - despite whatever different directors have done to the franchise, it's a mainstay and recognizable part of my existence.
In some ways, it is a buddy that will be by my side as I go through life.
Now, I can be cynical about it and see Bond as merely a corporate, franchised brand being foisted upon me as the unsuspecting consumer... But I choose to see Bond as something genuine people have put genuine care into creating for my consumption. It's not nearly so bad if you put it that way, eh?
I don't think Google is entirely off its rocker trying to portray itself as my "buddy" - because it is. It knows everything I've ever wondered about, either publicly or privately, and so far anyway its products respond to my needs much better than any of its competitors (GMaps vs. MapQuest, for example).
Since my comment is stuck in moderation limbo on reputationdefenderblog, I'll paste it here:
",,,I think Google meant the ad to show that it is powerful."
I don’t think the commercial’s primary intent was to display the "power" of Google, but rather the day to day usefulness of their engine.
Interestingly enough, the overwhelming initial response to this ad campaign has been positive from girls in their mid-20s (I saw a copious amount of "awws" in my twitter feed), and "mehs" from their male geek counterparts (who already probably use Google anyways.)
We weren’t exactly the target demographic of this ad. Home run, Google- you slammed this one out of the park.
I was initially going to write "Also: that geeks are much more numerous now", and on reflection that may've been a better observation.
Yes, Google users are geeks. Think back to the 1980s, or even the early 1990s. If anyone had suggested then that you type a few words into a box onto your computer screen to ask how to impress a girl, they'd be branded with "OMG you are such a geek." Except that "OMG" was itself a geek thing then. More likely, they'd be punched in the face.
It's amazing how standards have changed in 15 years, and how what was dorky then is normal now, and what was normal then is hopelessly out of date now.
This may be what the person responding to you meant. What I meant was that geeks as in privacy geeks have a mentality (as displayed in this article) that is far different than what a normal person feels. No one I was with said anything about finding the ad creepy, and many of them said it was sweet. All I meant to say was that this article just demonstrates that you should let a guy like it's author be in charge of your marketing, unless people like the author are the only customers you wish to attract.
25 comments
[ 7.5 ms ] story [ 431 ms ] threadDoesn't Google store search data even after the user has cleared it from their history?
Here's a thought: If you don't want Google to know what questions you ask - don't ask Google any questions.
Substitute %s with your search term.
If you move (change IP) Google will be able to track you if you use any of their sign-in services (example@gmail.com now regularly logs in from this IP instead of the old one). Your only chance of breaking this chain is either to regularly use proxies or to change IP and gmail (and other services) simultaneously.
Did you see the analysis of the AOL data leak? There is one particular woman in there where consecutive queries and a little editing show a story of her relationship with an abusive boyfriend, accidental pregnancy, abortion, and remorse.
People found it by, essentially, just grepping a flat text data file for terms they knew were likely to resolve to a hot button issue, then separating the wheat from the chaff. This is the "solution you can knock out in vi in ten minutes" answer. If you have a team of Google-style engineers, a large volume of search queries with unique IDs on them, and a Hadoop cluster available, you can make sift-and-winnow as many gigabytes as you have down to as few salient details as you want.
Trivial example: I would stake money on being able to correctly identify the gender of over 80% of users given 100 searches from them, and I would be able to do this with less than one day of programming. If you let me do it in other than realtime, I would be able to do this at terabyte scales using EC2 and Hadoop. And I'm just a guy who did natural language processing in college and has a bit of basic web-fu available. Think of what you could do with a talented engineering team, deviousness, and a few months to work on it.
Most of them in my case are software development-related stuff. Before being so afraid of “the power that google has” just check you history: it’s less frightening than what commercial made you believe.
But (almost) any publicity is good publicity, and this may inspire a few people to fly to Paris / get married in Paris / make a DIY crib etc, which has positive flow on.
To me it comes across as a mixture of arrogance and over familiarity. Google wants to know us better so they can serve more targeted ads. That desire is not reciprocal. Not for me at least. Its like it never crossed their mind that people may not be as pleased as they are at the idea of an ad network being their best buddy.
But often they are. Example: I grew up watching James Bond movies - I still watch James Bond movies, and there will probably be many more Bond movies between now and when I die. It's a known quantity - despite whatever different directors have done to the franchise, it's a mainstay and recognizable part of my existence.
In some ways, it is a buddy that will be by my side as I go through life.
Now, I can be cynical about it and see Bond as merely a corporate, franchised brand being foisted upon me as the unsuspecting consumer... But I choose to see Bond as something genuine people have put genuine care into creating for my consumption. It's not nearly so bad if you put it that way, eh?
I don't think Google is entirely off its rocker trying to portray itself as my "buddy" - because it is. It knows everything I've ever wondered about, either publicly or privately, and so far anyway its products respond to my needs much better than any of its competitors (GMaps vs. MapQuest, for example).
",,,I think Google meant the ad to show that it is powerful."
I don’t think the commercial’s primary intent was to display the "power" of Google, but rather the day to day usefulness of their engine.
Interestingly enough, the overwhelming initial response to this ad campaign has been positive from girls in their mid-20s (I saw a copious amount of "awws" in my twitter feed), and "mehs" from their male geek counterparts (who already probably use Google anyways.)
We weren’t exactly the target demographic of this ad. Home run, Google- you slammed this one out of the park.
Yes, Google users are geeks. Think back to the 1980s, or even the early 1990s. If anyone had suggested then that you type a few words into a box onto your computer screen to ask how to impress a girl, they'd be branded with "OMG you are such a geek." Except that "OMG" was itself a geek thing then. More likely, they'd be punched in the face.
It's amazing how standards have changed in 15 years, and how what was dorky then is normal now, and what was normal then is hopelessly out of date now.
1. The ad wasn't made for the Super Bowl
2. There are other ads in the "search stories" series