The Rise and Dangers of Big Data & “Reputation Scoring”
“Big Data ” – the term referring to the unprecedented volume of personal information being created online every minute – is being vacuumed up at blinding speed by companies that specialize in finding, organizing and selling it. They sell it to employers, job recruiters, insurers and in some cases to anyone with fifteen bucks and an interest in your personal life. This includes companies you know well, like Nielsen, the media-research company best known for tracking our viewing habits, and companies you’ve never heard of, like eBureau. eBureau has “….billions of records across thousands of databases that cover nearly all US adults and households…[and it] adds over 3 billion new records each month.” (This means they add one hundred million records a day.) It uses formulas to assign you an “eScore.” It’s like a credit score, except it’s not. The eScore, already in wide use, is based on new-media data such as your social-media postings, blogposts, and records of online purchases. Their use does not trigger any of the legal protections associated with traditional credit scores like FICO. In other words, you have no general legal right to know your eScore, to demand that it be accurate, or to demand that errors be corrected. Despite this, your eScore is used by businesses to decide whether to hire you, insure you, promote you or to do business with you at all. Some believe it was specifically created to allow businesses to score you without the hassle of credit-disclosure laws. You’re still being scored. You just have fewer and, in some cases, no rights.
Think of your eScore as your “reputation score.” It’s Big Data’s end-of-the-day, bottom-line on who you really are – when everything imaginable has been blended, shaken, stirred and baked and, finally, reduced to a single number.
To create these scores, data-mining companies download and permanently store everything their computers can grab. They then process it using mathematical algorithms to assign numerical values to the data, based on the data’s relative importance in defining key characteristics about you: reliability, honesty; likelihood of repaying debt, and so on. Quantities equivalent to the entire Library of Congress a hundred times over are being downloaded from the Internet every day. Companies download and store data even if they don’t know what to do with it. Storing data costs almost nothing. You can store a terabyte of data – roughly 75,000,000 pages’ worth – on Amazon’s S3 servers for about $10 a month. So it’s less expensive to grab everything and keep it forever than it is to pick and choose what to keep and what to delete. The real effort is devoted to organizing the data for sale. It’s used in ways both creepy and surprising. How creepy? Nielsen got caught copying all the private message boards on PatientsLikeMe.com, a site for people with serious illnesses to share their stories and talk openly about it. And how surprising? eBureau scores are used by companies to decide how long you’ll be on hold. Callers with high eScores that are most likely to buy are taken off hold very quickly; complainers and perceived time-wasters with lower eScore are routed to call-overflow centers and face lengthy hold times. This routing calculation takes just a fraction of a second for the computers to determine. (This is why most hold announcements no longer tell you how many callers are ahead of you; instead, they just tell you how long you’re going to be on hold.) If you still thought your hold time was based on the order in which you called, you should read more about Big Data.
Another company, FirstAdvantage, sells a product called Esteem that lists employees who’ve admitted stealing from their workplace. Who uses Esteem to decline job candidates? Target, CVS and Family Dollar, among others. This is a terrible product. I’ve known hourly-wage employees to admit modest thefts because they were pressured by loss prevention employees eager to resolve a matter. The employees were led to believe they could keep their low-paying jobs and continue supporting their families if they just admitted guilt, even if they were innocent. And no one’s told an admission will wind up in a commercial database marking them as an unemployable thief.
Finally there are companies that vacuum up, organize and sell your data to anyone. Spokeo sells your name, address and phone, and it can also tie you to usernames you’ve used on social networks, blogs, photo albums, dating sites, ecommerce sites, and other public web services. Do you use the same “anonymous” username across platforms? It may be unwise.
8 Ways to Protect Your Job Searches and Career Opportunities From Big Data & “Reputation Scoring”
It’s important to appreciate that reputational scoring often does its damage behind the scenes. You will rarely discover you lost an interview or job (or insurance, or a loan) because of Big Data. You cannot wait until after the lost opportunities to confront this threat. You must be proactive, by thinking ahead about your online activities on any site. Great people have been losing jobs for years now because of online activities, and it’s going to get worse.
1. Be absolutely consistent when posting credentials and resumes online, whether on job-searching sites or on social media. Reputation-scoring systems may flag typographical errors as “possible resume fraud” when you post a corrected, but different, version. Did you list a B.A. degree by mistake when you mean B.S.? Do you use customized resumes that list different work histories depending on the position, with a little bit of fudge to cover time gaps resulting from omitted positions? Are the dates of employment inaccurate? Do your past job titles match from posting to posting? Computers are literal. They will red-flag all variations in your claimed credentials, regardless of the good-faith explanations you might have had.
2. Lock your personal stuff down online. If you want to post purely social, personal data, set your profiles to private. If you can’t do that, use unique usernames on each that prevent algorithms from spotting subtle similarities and linking them to your real identity.
3. Refrain from discussing medical issues online. Data-miners do brisk business with insurance companies. Posts about medical conditions might wind up costing you insurance coverage. Or, they might deny you benefits later. Many insurers will refuse to pay benefits if they can show you had health conditions that you failed to disclose on the application for benefits. Facebook is a prime source for catching people talking when they think no one’s looking.
4. Watch out for who you use as references. Some employers check the references of your references. Reputational scores may be weighted in part on the scores of your listed references or even your profile friends. If your references score poorly, or have lengthy criminal histories, your reputational score may be affected. Remember the classic Framingham Heart Study, which found that a person’s eating habits, good or bad, tended to mirror that of their acquaintances? There is considerable evidence that behaviors spread socially. Reputational scoring may adjust your score if a significant percentage of your references, friends lists or neighborhoods have histories of criminal convictions or bankruptcies for example. Remember your mother’s warning that if you hung around bad people, you’d become one of them? Who knew? Moms are the ultimate reputational scorers, aren’t they.
5. Be mindful of online posts through sites like TripAdvisor or Yelp. A history of highly-critical posts about hotels, restaurants and other establishments may affect your ratings. Use entirely anonymous usernames if you must, and don’t use the same one all the time. The algorithms used to link you to postings use tens of dozens of evaluative criteria. If they’re not sure, they may include the questionable data anyway and simply mark it as unclear. Even a faint aroma of misconduct might disqualify you from consideration for a job.
6. Be careful about admissions of misconduct on the job or online. Social media posts about what you took from the workplace, or your neighbor’s shed, or a store will come back to haunt you.
7. Assume that photos of you online that contain your face will be included by data miners in their reports. They search across the obvious sites, like FaceBook and Linkedin, but go far beyond that and may include postings and photos on Flixter (your movie ratings), Pandora (your music play lists), Flickr/Picasa/Smugmug (your personal photos), Twitter, Digg, StumbleUpon, Craigslist, Amazon wishlist items and many more.
8. See a lawyer immediately if you discover a report contains false or inaccurate information. Many Big Data brokers cross the line in gathering information, providing it, and in selling it as being more accurate than it is. There are a range of claims that may be available to you, including defamation and invasion of privacy.
Special Bonus Material: As part of my research for this article, I subscribed to a data-mining site available to the general public and searched my own name. Here are the photos it retrieved, all marked as nothing more than “possible.” Some pictures included in my profile linked to “James Garrity Tallahassee” are comical. (I assure you, Mr. Recruiter: I have never been, and I have never associated with, a towel-wrapped angry Cocker Spaniel.) But remember this: If the pictures are this far off, how far off is the data? And if the data is just as inaccurate, what are the odds that you’re going to be passed over for the very best positions?
Categories: Background Checks, Bad/Retaliatory Job References, Credit Checks by Employers, Criminal History Checks by Employers, Defamation, Fair Credit Report Act, Internet Usage, Invasion of Privacy, Job Searches & Interviews, Off-The-Job Conduct, Privacy Violations, References, Slander