20 million reasons to cancel AOL

Update: You can now search the AOL data from your web browser.
As promised earlier, I did some scans through the massive privacy invasion from AOL for some real state search insight. I’ll leave it to other sites to search for the tell you about the disgusting things people search for.
Not many AOL searchers are looking for “seattle real estate” in those words – in fact only 21 of the 20 million queries contained that text and those users largely went to the big Google-optimized sites like Seattle Power Search (the number one result) or the Seattle Times (the number 3 result).

AOL users found Rain City Guide through many long-tailish routes, with relevant keywords like “zilllow” (sic), “small condos,” “seattle real estate,” “earnest money recipt” (sic), and “five factors that determine if an idea is a good investment opportunity.” Guilty-conscience user 917673 came to us while searching for “sellers disclosure for condominium complex.” And User 1636230, who came to Rain City Guide after searching for “seattle real estate,” also visited Winderemere and Seattle Power Search. One searcher also found us when they searched for “dustin dustin.”
Here’s where this data goes beyond our own site (and where it gets creepier): what preceded and followed those inquisitive searches? Can we tell something about these people? Well, User 1636230’s interest in real estate was passing. They searched around for five minutes in March:

  • www.happydogtoys.com 2006-03-21 14:12:20 2 http://www.happydogtoys.com
  • www.happydogtoys.com 2006-03-21 14:12:20 2 http://www.happydogtoys.com
  • hometown realty executives 2006-03-21 15:30:32
  • hometown realty executives seattle 2006-03-21 15:30:41
  • hometown r.e. executives seattle 2006-03-21 15:31:05
  • hometown real estate executives seattle 2006-03-21 15:31:24
  • seattle real estate 2006-03-21 15:33:18 1 http://www.seattlepowersearch.com
  • seattle real estate 2006-03-21 15:33:18 8 http://raincityguide.com
  • seattle real estate 2006-03-21 15:33:18 9 http://www.windermere.com
  • locks for love 2006-03-22 10:52:35 1 http://www.locksoflove.org

Then decided to look into building their own home a week later:

  • lux homes 2006-03-28 15:33:40 1 http://www.luxhomesllc.com
  • woodenville builders 2006-03-28 15:34:56 (4 more identical searches)

User-1636230 then went on to search for approximately 10,000 pet related items and for much sadder subjects, including for cancer drugs and incontinence.

What of guilty-conscience-User 917673? They were clearly concerned about their condo and they didn’t want to tell the buyer. Here are three of their searches (of over forty):

  • condominium disclosure by seller in los angeles
  • arbitration for selling or buying a condo
  • consequences of no disclosure from seller

Sounds like someone got a bum deal on that condo!

In looking at the other Seattle Real Estate searches, it seems that the adage that buyers and sellers go with the first agent they talk to does not apply to searches (no big surprise here). Searchers go all over the internet and leave and come back to the same search repeatedly. If they’re as committed as User 917673, they use lots of slightly different word combinations. What I found interesting was watching users hit a site that they are interested in, then go on to search for that company’s or person’s name to see if they can find some background (so it is good to be on a first name basis with the search engines).

I only found 668 occurrences of “cancel AOL.” I suspect and hope there will be a lot more this week.

Real estate search patterns and AOL users

Yesterday AOL proudly announced the release of 20 million web queries from 650,000 users (screenshot), with each user “anonymized,” but identified by a unique ID. This is appalling – it means that potentially thousands of social security numbers and email addresses are now free for spammers and thieves to harvest, along with a lot of other personally identifying information. Think about what you search for – email addresses, people’s addresses, business secrets and even social security numbers come to mind. AOL quickly realized their mistake and pulled the plug, but not before the dataset had taken on a life of its own.

So, spammers and thieves are having a field day, but now that it’s out, we might as well use it for educational purposes. It’s a big, unwieldy file, but I’ll try to post some real estate search patterns by tomorrow. If you’re hoping to do your own analysis on this dataset, I wager that there will be a nice web interface for you to use within a week (Consumerist thinks so too). I’ll let you know when it pops up.
More on the ramifications of the release at TechCrunch. If you’re going to cancel your AOL account, good luck.