A few months ago, LinkedIn came to me with an unusual proposition. How would I like to meet my future self?
The offer came a few weeks after LinkedIn opened its publishing platform, meaning anyone on the service could write a blog post tied to their profile. I was looking for something personal to write about — and that's when company reps suggested they could identify a user who fit my projected career path, predicting where I might be professionally in five years.
Thus was born Future Kurt, a first for LinkedIn. The company has never forecasted on behalf of a user before. Its only stipulation — that I make sure my profile was up to date. LinkedIn would do the rest.
The man responsible for finding "Future Kurt" was Lutz Finger, LinkedIn's director of data science & engineering. Finger, a German-born quantum physicist who also happens to be a Fulbright Scholar, has managed LinkedIn's 18-person data science team since last October.
Despite his job title, Finger doesn't have a fondness for big data. The term has become something of a buzzword in the tech world — but doesn't provide any answers, Finger says, unless you know how how to analyze it.
“Big data is a headache," he said. “Big data is a lot of hype. You don’t want big data. You want one name."
For Finger and his team, sifting through LinkedIn's copious amounts of user information is best simplified into two things: a question and the measurements needed to answer that question.
Finger's process for finding Future Kurt required running LinkedIn's 300 million member user base through a series of relevant filters (the measurements). The challenge was identifying which filters to use. Did it matter what college I went to? What about my gender? Which experiences from my profile best predict where I will be five years down the road?
Here's how LinkedIn tackled the problem.
The Process
Step One. Future Kurt Pool: 300 Million
Starting with LinkedIn's total user base of 300 million, Finger first identified the pool of users who were once journalists — this being the only profession I've ever had. Finger entered keyword searches like "writer," "editor," and "journalist," meaning the system would also find variations such as "senior writer" or "freelance journalist." He also kept the search relatively simple, to profiles written in English.
Step Two. Future Kurt Pool: 60,000
Next, Finger and his team examined the length of the average journalism career. It turns out, journalists don't stick around for long.
In fact, Finger found that 40% of journalists are out of the profession by the end of their third year. I'm in my third year now. By year eight — five years down the road — 70% of journalists have switched professions.
This turnover rate startled Finger. As he put it, "Is our methodology wrong, or is it really that sh*tty of a job?” He compared the numbers with turnover figures for a career in law. By the end of year six, 50% of law professionals switch careers. (They likely stay longer due to the high costs and commitments that come with law school.)
Finger concluded that his methodology was correct.
The data claims I'll be out of journalism by year eight, so Finger and his team narrowed the search by identifying users who left the journalism field between years 3-8 in their career. Future Kurt would not be a journalist.
This may be a flaw in the system — I have no intentions of leaving journalism, and the data certainly didn't coincide with my own five-year plan. Then again, maybe other third-year journalists felt the same way before switching careers. Let's see where this is going, I thought.
This is how the 19,000 remaining non-journalists break down by career:
Step Three. Future Kurt Pool: 19,000
Here's where more specific filters came into play. Finger and his team combed my profile to determine which career events and skills were most relevant to Future Kurt, and came up with five main criteria:
- Some involvement with social media (my current beat)
- Worked for a tech publication
- Worked in business news
- Completed an internship
- Skills match (50%)
Flags like university and gender were ignored; they aren't indicative of my professional outcome, Finger says.
A total of 939 LinkedIn users shared one commonality with me — that is, either a 50% skill match or one of the other four criteria Finger identified.
More importantly, 24 LinkedIn members shared two commonalities. The pool was narrowed once more.
Step Four. Future Kurt Pool: 24
With just 24 candidates remaining, Finger decided data had taken them far enough. Humans took over. A portion of LinkedIn's public relations team took the information they knew about me from our frequent interactions, and boiled the list down themselves.
They looked at information such as location, interests, and areas of study. The group looked specifically for a candidate who worked in sports, a passion of mine — but unfortunately, no one fit the bill. Still, they narrowed the pool to three names.
Step Five. Future Kurt Pool: 3
With three candidates remaining, I was asked to pick the one I identified most closely with. The LinkedIn data machine didn't pop out just one name as I'd hoped, but the strategy made sense — who knows the vision I have for myself better than I do?
Each of these users was quite literally one in 100 million, so they all had familiar backgrounds and skill sets. All of them created content, just not for a news outlet.
One candidate was a social media manager for a major tech company. Anther designed in-house reports and presentations at an architectural consulting firm.
The third candidate stood out among the pool. I found Future Kurt. And he was a she.
Interviewing Future Kurt
Mussarat Bata runs content marketing and social media strategy at SBE Entertainment, a hospitality company in Los Angeles.
She got her first full-time job at Time Inc. with People Magazine (I spent my first year at Fortune, also owned by Time Inc.); she wrote for UCLA's university paper, The Daily Bruin (I was Editor-in-Chief at Santa Clara University's paper, The Santa Clara); she even studied political science (so did I) and we have 12 skills in common.
These similarities were not the reason I selected Bata from LinkedIn's final list. Instead, I took a guess as to why she ultimately left journalism: to pursue her passion for fashion and beauty. I assumed that if I ever left journalism the way LinkedIn believes I will, it will be to follow some kind of career in sports.
LinkedIn contacted Bata — who was surprised to hear about her "past self", but found the idea interesting. "It feels weirdly futuristic, but in a good way," she said. "With technology like this comes more awareness and perspective about the job market, and I feel like that's really needed, especially for our generation."
It turns out I had things backwards. Fashion hadn't driven Bata out of journalism; instead, it was the reason she got into journalism in the first place. “I was able to really pursue a passion I was very very interested in," she told me. "Then I saw an opportunity to make that a career through journalism.”
I got my journalism start in similar fashion. I began writing after I realized my dream of playing professional baseball was slightly out of reach — apparently becoming a professional athlete requires physical skills beyond "trying really hard." I started, instead, as a sports reporter at The Santa Clara and never turned back.
Bata jumped out of journalism for the same reasons thousands of other journalists have switched careers: a lack of jobs and a lack of money to support those who do remain.Bata jumped out of journalism for the same reasons thousands of other journalists have switched careers: a lack of jobs and a lack of money to support those who do remain. "I have friends who I went to journalism school with who are still toughing it out," says Bata. "Four years later, they're still living paycheck to paycheck in New York City. As much as I respect the art of journalism and the craft, I just couldn't live like that."
Bata has no regrets about her career switch, and started her post at SBE a few weeks back. It's no surprise to Bata, though, that LinkedIn predicts I'll be out of the profession in the next five years. She noticed the trend before she even completed her master's degree.
"When I was in journalism school [in 2009], everybody who I was in school with was like, ‘well maybe we can do PR, maybe we can do other things,'" she recalls. "We’d already given up on the idealistic world of being journalists.”
I would venture that most journalists have heard similar frustrations from former colleagues. The interview didn't paint a promising picture for the industry — but then again, I've already witnessed issues of job scarcity. This wasn't news.
Five years ago, however, when I was busy scraping together stories for a small college newspaper, looking forward five years in this way might have had a major impact on my career — especially if it put me in touch with someone who would be, in effect, a mentor figure. It's the kind of service you can imagine LinkedIn offering to college students.
Finding Future You
It's likely that one day, LinkedIn will offer a feature like this for all of its 300 million members. Imagine a tool that takes your profile data and returns a dozen potential mentors in seconds.
It's possible to do the digging yourself using LinkedIn's search features, but an automated tool from LinkedIn would certainly be more efficient. A LinkedIn spokesperson says that the company isn't building out the feature yet, but hinted at possible interest.
“Students are our fastest growing segment," Joe Roualdes, a LinkedIn spokesperson, told Mashable. "Being able to show them things that they can do to advance their careers — go to the right college, be successful, get a certain job, drive a certain career outcome — that obviously interests us.”
Based on my experience, it doesn't appear LinkedIn is close to finalizing a tool for the general user base — not yet. An automated system doesn't take into account external factors, like desired job location or personal relationships. Plus, finding Future Kurt took weeks — although to be fair, finding him was hardly a top priority for the company's data science team.
Finger claims he could recreate the search in under an hour, largely because the framework is now established.
LinkedIn does have its eyes on student users. As Roualdes mentioned, it's the company's fastest growing user segment, and LinkedIn unveiled University Pages in August to help college better connect with current and prospective students. Last January, LinkedIn unveiled an alumni search tool so students could find others from their university who took a specific career path.
It's an interesting opportunity for LinkedIn, a chance to move from supporting your career to actually predicting it. It's not yet a perfect system, but if you ever get the chance to interview your future self, I highly recommend it.
You might be surprised by what you have to say.
Have something to add to this story? Share it in the comments.
Powered by Livefyre