One of the most important areas within HR analytics is employee retention, with a variety of products in the market that use predictive modeling to help companies determine which valuable workers may be at risk of leaving the company. A startup called Joberate is putting a new twist on this by mining workers' public social media accounts, as the Washington Post reports:
Most people looking for a new job -- at least if they currently have one -- use their personal email to correspond with a prospective employer. They don't tell the people they work with they're being recruited. They slip on a suit jacket for the interview after leaving the office building. In other words, they carry out the process in secret.
Or so they think. A start-up that tracks an individual's job search activity in their public social media accounts is quietly -- and some would say creepily -- calculating a score it says helps represent how likely each one is to be looking for a job.
The start-up, Joberate, scrapes publicly available data from millions of individuals' online social media accounts, or buys it from other parties, to assign what it calls a "J-Score" that estimates their level of job search activity, likening it to a FICO score. If the person starts following company accounts on Twitter, clicks through to articles about résumé writing or career-related content in their Facebook feed, or begins making a bunch of professional connections on LinkedIn, their score goes up. Joberate then shares these scores with clients -- typically to help employers keep tabs on talented outsiders or see how engaged their own workers are in their jobs.
Joberate doesn't see activity that has been set to private, says co-founder and chief executive Michael Beygelman, and it doesn't pick up on general online searches outside social media accounts or most activity on job boards. But if you use an "Apply with LinkedIn" button on a job posting, comment on a story about job searches that uses Facebook to collect comments, or use some niche industry sites such as Stack Overflow that combine aspects of social media and job boards, Joberate could detect your activity and include it as part of your score.
Beygelman told the Post that Joberate has clients among 14 of the Fortune 100, but that most of them use the tool at a high level, looking for trends within departments.
There's nonetheless cause for privacy concerns with tools that take this approach to data collection, says Constellation Research VP and principal analyst Steve Wilson.
"Joberate is said to scrape data from the public domain to work out what people are thinking of doing with their careers but I am here to tell you the 'public domain' is one of the great red herrings in privacy," says Wilson, who leads Constellation's research into privacy and security matters.
"There is this concept that's emerging of 'collection by creation,'" Wilson adds. "If an algorithm creates fresh Personally Identifiable Information (PII) from raw data—like your LinkedIn updates, or your shopping history—then that algorithm has collected PII as a result. And whoever is running the algorithm is exactly as liable for the PII they are creating and collecting as if they had gone around with a clipboard asking people face to face what their career intentions are. Think about it. Big data is often used as a way of avoiding face-to-face data collection."
Well that's not just creepy: It's a clear cut breach of the Collection Limitation Principle in dozens and dozens of jusridictions worldwide. Businesses are not allowed to collect any old PII they like, without good cause or without consent."
Constellation Research has been working for years on what Big Data means for "Big Privacy," Wilson notes:
- Don't use Big Data algorithms behind peoples' backs. If there is a win-win in data mining, then try to bring customers into the bargain. Seek their consent for what you want to do with their data, even if—nay, especially if—the raw data is harvested from the public domain.
- Write a "Big Privacy Policy" that explains why you collect data, what you know about people, and what you might know abouyt them through Big Data. Give people visibility of how PII out them flows, and some control over that.
"We encourage businesses to be proactive about Big Data privacy and to innovate in their processes and customer relations," Wilson says. "Because if they don't, they will get caught by surprise by the steadily evolving international privacy jurisprudence." To wite:
- Collection by Creation is expressly recognized now in Australia by the Deferal Privacy Commissioner.
- The new US National Privacy Research Strategy defines privacy to explicitly include the creation of PII: "privacy concerns the proper and responsible collection, creation, use, processing, sharing, transfer, disclosure, storage, security, retention, and disposal of information about people."
All that said, we're in early days when it comes to Big Data and the career management industry, Wilson says, pointing to Microsoft's plans for the rich trove of information held in LinkedIn. But it's all about doing it right.
"I think we can trust Microsoft to use LinkedIn's graphs with sensitivity and respect, based on the company's long term privacy efforts, their original Differential Privacy research, and their international business which needs to accommodate the strict application of the Collection Limitation Principle in Europe."
24/7 Access to Constellation Insights
Subscribe today for unrestricted access to expert analyst views on breaking news.