The most dangerous thing about all the tracking that’s done on us over the internet is not how much computers get to know about us, but how wrong they can be.

That presents a bunch of worries over those transactions (such as credit checks) that rely on the automated crunching of mounds of data.

But if you need to know a lot of detail about a person with a high degree of certainty, data mining isn’t where the gold is. For a good human investigator, modern computing brings to mind the primitive clunker pictured below when compared to a keen mind and an impressive bit of technology known as the telephone.


Take the latest, well-researched treatment of the subject by Joel Stein in TIME. With all of the ingenious tracking technology that follows him around on the internet, data miner RapLeaf thinks he has no kids (when he has), works as a medical professional (which he isn’t), and drives a truck (which he doesn’t). Google Ads and other well known data accumulators also get him very wrong.

His conclusion is that “RapLeaf clearly does not read my column in TIME.” My conclusion is that RapLeaf could be outsmarted by any decent investigator who can read and then substantiate Stein’s writings by chatting with a few people who work with him.

Surf over to, and one of the data miners might conclude you’re a Second Amendment enthusiast when the truth is, you were researching a case about guns. The best way to find out someone’s feelings about guns is to see what they’ve written in the past through a thorough press search, check their political donations, and then to call people who knew or know them.

Maybe such an error is costless for guns people who want to send you a virtually free spam email, but for someone thinking of hiring you or investing with you, which places you visit on the internet is hardly a reliable indicator to color such an important transaction.

Data mining sounds creepy. One of the leading journals in the field features a current article called Limitations of Matrix Completion via Trace Norm Minimization. It sounds as if it’s beyond the mathematical skills of most of us and probably is, but so what? The best math modelers failed to spot the various financial bubbles that have burst over the past decade. Nobel laureates in math and statistics have helped with hedge funds that blew up from too much risk taken on. It’s all laid out beautifully in Nassim Taleb’s The Black Swan.

Besides, Stein makes the point that our identities have never been completely within our control.

“Our friends keep letters we’ve forgotten writing, our enemies tell stories about us we remember differently, our yearbook photos are in way too many people’s houses. Opting out of all those interactions is opting out of society.”

What are some of the bits of information a good investigator has always been able to get about most people, and still can?

  • How much you paid for your house, who you bought it from, and how much your mortgage is
  • Which political candidates you give money to
  • Whether you’ve ever been suspended or disciplined from your profession or occupation that’s subject to a state license
  • What kinds of equipment and car leases you have
  • What side companies you run out of your house or office

You can try finding this out on the web. Sometimes you’ll succeed and sometimes you won’t. But hire an experienced fact investigator with a computer and telephone, and you will be able to get most of this information in a completely legal, ethical manner.

Just as you could have done 20, 30 or 40 years before Google and modern data mining were even invented.