Two Teens at 26?! What Big Data Got Wrong

    Slide Show

    Overstepping the Line Between Big Data and Big Brother?

    I’m sure you’ve heard about Big Data’s potential to revolutionize business, including the assertion that it will give you more insight into customer behavior. Statistically, that may be true, but a recent CNN Money article challenges the idea that Big Data will allow organizations to “know everything” about us as individuals.

    Using Acxiom’s beta website,, CNN Money ran an informal survey of 10 people (yes, I know). Acxiom, by the way, says its massive database draws from public records, surveys and other data sources to cull information such as education, political leanings, ethnicity and income level on 190 million consumers.

    But when the CNN volunteers checked their profiles, each found “at least one major inaccuracy” in their profile, the article notes. One example: The database informed the writer she had two teens, at 26—“which is just about biologically impossible,” according to her. (Technically, it’s not biologically impossible, but the point is she doesn’t have any kids, much less two teenagers.)

    Of course, anecdote is not the plural of datum, and a survey of 10 from 190 million is hardly representative. But, heck, if you’re a data person (and you probably are if you’re reading this), you understand that better than anybody.

    Still, the article is a fun read, although judging from the comments, it will do little to cool down the complete freak-out people often have when they learn about Big Data. In fact, for some, it seems to have been an eye opener.

    Acxiom doesn’t claim 100 percent accuracy— and who could? On the contrary, the company warns that up to 30 percent of a person’s profile information may be wrong at any given time.

    But I see two really interesting takeaways about this experiment and Acxiom:

    1. What the data got wrong: The CNN team found the database was more likely to be on target about their interests than other key demographic data, such as marriage status and whether or not the person had children. From a business perspective, that’s puzzling news. On one hand, you know their interests—hooray, mission accomplished! On the other hand, you may have just sent out 2 million direct mail pitches on baby strollers to people who don’t have kids.
    2. Acxiom does something the experts keep telling us to do: It allows the consumers to be data stewards for their own information, giving them an option to correct their profile, but also to opt out or suppress some of the data. The sneaky part is, as one reader points out, to do that, you have to fill out an information form, thus perhaps unintentionally improving the accuracy of their data.

    Definitely check out the article. It’s interesting to read the individual experiences with the data. I find the comments equally fascinating, but that’s probably just my weird attraction to how extreme Internet comments can be.


    Tuesday, Sept. 10, 9 a.m. PT. “Harnessing the Power of Big Data for Healthcare Organizations” by TDWI. David Stodder, research director for BI at The Data Warehouse Institute, will drill down on the technology tools health care needs to leverage Big Data for better customer care, managing costs and more efficient operations.

    Tuesday, Sept. 24, 1 p.m. PDT. “How to Get Rolling with Big Data Analytics.” As much as we want Big Data to be simple, it’s still different from traditional enterprise data systems and will require new skill sets, even for granular integration work, according to the Bloor Group. IT analyst Dr. Robin Bloor will talk about the problems with modern data volumes and types in this free event. IBM’s Rick Clements will talk about the company’s Big Data platform, InfoSphere and share Big Data use cases.

    Loraine Lawson
    Loraine Lawson
    Loraine Lawson is a freelance writer specializing in technology and business issues, including integration, health care IT, cloud and Big Data.

    Get the Free Newsletter!

    Subscribe to Daily Tech Insider for top news, trends, and analysis.

    Latest Articles