NSA Surveillance: Metadata Isn't What You Think It Is

We’ve been going to great pains to discuss our government’s collection and use of something called “metadata” over the past couple of weeks. Whether you are of the opinion that all the current hyperventilation is too little too late and we ought to have been much more cognizant of these things back in 2001, when the Bush administration invented the Department of Homeland Security and The Patriot Act, or whether you don’t remember because you’re too young or too distracted to do so — it is now time to understand the cause of all this ruckus.

In 1923, a brilliant young man named Arthur C. Nielsen began a company that achieved a huge amount of both subsequent fame and notoriety in the marketing research world as ACNielsen and, finally, as The Nielsen Company. In the days before computers, all consumer research/inventory change data was recorded by hand on paper, by Nielsen representatives. If you ever wondered where the sobriquet “beancounter” came from, that is its origin. It is also mine — I joined ACNielsen in 1977 as a representative, with a BA in English literature and a bright future.

I progressed up the corporate ladder as the company transitioned from its paper-data phase (imagine stacks of reports higher than you are tall, delivered to your office every month) to the computer era, wherein not only was more data storage possible — it was also possible to manipulate and analyze it in new ways.

Enter the geodemographics geniuses. At first, geodemographics married census data at Zip code or census-block level with a geographic radius, to determine the makeup of retail neighborhoods for the most profitable location of such outlets as grocery stores, fast food chains, box stores, etc. Its use as a real-estate location tool was soon outstripped by the enhancement of socioeconomic segmentation/clustering tools, which used the consumer purchasing information ACNielsen collected to create profiles of the “people most likely to…” Once those were linked with geodemographics, it became possible to target markets precisely to neighborhoods, down to an individual apartment building or suburban block.

It was only a matter of time before the politicians got into that act. Oddly enough, the Republicans caught onto targeted mailings before the Democrats did. That is one of the reasons for their electoral successes in the 1980s and early 1990s.

Once political campaigns were on the bandwagon, of course the functioning government would eventually hop on, too. Between the mid-'90s and 2001 — essentially the end of the Clinton administration and the beginning of the Bush administration — only some governmental agencies upgraded their computer systems and took advantage of the new tools. Once Congress passed the PATRIOT Act, however...

So, all that census data taken since the first census in 1790 has been computerized, down to the Zip+4 level. Individual households can be targeted precisely via geodemographics. They can be profiled and classified via the socioeconomic segmentation provided by consumer marketing data, and they can be located.

With the advent of cellphones, smartphones, GPS systems, iPads, and WiFi, the picture is suddenly in motion. Not only is it possible to know preferences of individual consumers, attached to socioeconomic data and residential location, now it’s possible to know this on the fly. There are drawbacks to “living out loud.”

Before you panic, please understand what datamining is. That is the process of sifting through metadata — via sophisticated mathematical programs called algorithms — to find patterns that predict certain outcomes, as expressed by a coefficient of probability. When President Obama describes this as “looking for patterns,” he’s simplifying enough for a five-year-old to understand, and in the process minimizing the importance of the mathematics involved. This stuff is way more complex than the storm-tracking models for hurricanes that you see on The Weather Channel, as well as way more precise.

Thus, if an algorithm instructs the computer at NSA to search for phone calls between Americans and Pakistanis, which use the words “jihad,” “bomb,” “terrorist act,” and/or “IED” – it is not going to pinpoint a call some Pakistani student in Detroit made to his grandmother in Islamabad to wish her Happy Birthday. It is not going to pinpoint anything, but it will isolate time, duration, and content, and flag those for further examination.

I’m not glad about all this metadata collection and have not been since the PATRIOT Act passed in 2001, but it’s not President Obama’s fault. The time to scream and holler about losing our civil liberties was a dozen years ago. If you want them back, elect some different representatives to Congress.