Several weeks ago I looked up a book on Amazon but I didn’t buy it there, I grabbed it at my local Barnes and Noble.  The other day I was on a website and saw that the ad on the side column was for the book I had and was reading.  I went to another website and there was the book advertisement again.  It followed me to three other websites and then I cleared my cache and cookies.
The more I thought about it, the more annoyed I got about how intrusive it was that advertising mines my search history. I brought it up to my husband and said he’s found it even worse.
Imagine going to a library and the first time you go in, you ask for 18th century literature.  The librarian takes you to the correct section.  Then you ask about Corvettes and she takes you to a section on ships. “No,” you say, “I’m looking up information on Corvettes.”
“Yes,” answers the librarian, “Corvettes in 18th century literature.”
“No,” you say, “I want information on Chevrolet Corvettes.”
The librarian smiles at you in a patronizing manner, “Silly person, cars did not exist in 18th century literature. No search results found.”
So you leave the library, frustrated.
As you’re in the parking lot you figure you will go back and find a different librarian but that doesn’t help; your record says you want 18th century literature.
So after leaving the library you send a letter asking for your records to be destroyed. Then the next time you go to the library you have no trouble getting the manual for the 67 Corvette.  However, once you get that manual, no one will let you into the section about Russian composers.  Once again, in order to get the information, you need to leave the library, erase your records and then go back.
Lately, this analogy has been my experience with many websites. I have diverse interests.  Data mining has trouble with that. My husband has run into a few retail sites that won’t show him the parts he needs for the current stereo he’s working on because his order history has him owning a different brand and model.  My husband’s hobby is refurbishing stereos and makes and models rarely stay around the house. So big data remembering what he owns is more a problem than a solution.
I wish there were a place that consolidated all my hobby purchases. I can’t remember if I bought Nora Corbett’s Holly but I’m seeing the ads still.
I can see how retailers think that repeating the information from searches will bring people back to buy, but the only one that ever worked for me was when Think Geek sent a coupon for the items I had put in my shopping cart and left the site without finalizing the sale. However, now I always close the site and buy my stuff the next day. Sometimes I get a coupon sometimes I don’t. I’m surprised the bot doing the data mining hasn’t figured out this shopping pattern.
What big data mining things do you notice?