Playing for keeps with IT workers? Watch their online activity

Associate Professor of Information Systems John (Zhongju) Zhang recently correlated that taking part in virtual networks corresponds to both voluntary quits and internal promotions, and the way an employee participates foreshadows which of these results is more likely to happen.

A twitch, nervous laugh, or fidgety fumbling with a wedding ring may be classic tells at the poker table, but what giveaways signal that an employee may be soon trying his or her luck in a different job? There are plenty of well-known behaviors that might give you a clue. Less effort, grumbling, leaving early, and dressing a little nattier than usual are a few.

One behavior often overlooked is participation in online communities, the sites where tech workers convene by a screen to ask and answer work-related questions. Associate Professor of Information Systems John (Zhongju) Zhang recently correlated online participation in the SAP Community Network and subsequent job changes. He found that taking part in the online give-and-take of these communities corresponds to both voluntary quits and internal promotions. The way an employee participates foreshadows which of these results is more likely to happen.

Leaving a mark

What started Zhang’s research was a consulting gig for one of the top insurance companies worldwide. Managers were worried about employee retention, and well they should be. The Society for Human Resources Management estimates that the cost of replacing an employee ranges between 90 percent and 200 percent of the employee’s annual salary. Costs factored in include hiring and onboarding, as well as training, ramp time to peak productivity, higher business error rates and impacts to corporate culture.

In IT, turnover leads to knowledge loss, operations disruptions, and more, Zhang notes in a recent paper. “On average, an employer needs an employee to be with the company for five years to break even,” he says.

But that’s a hard sell in IT. A study by PayScale found that even well-respected tech giants like Google and Amazon have an average employee tenure hovering around one year. It’s not that the free cafeteria, onsite massage rooms, and nap pods fail to please. It’s that tech is highly competitive.

By some accounts, the average tech worker only stays on the job for three years.

That’s why Zhang’s consulting client wanted to look for other causes. “There is a lot of literature on employee satisfaction and turnover,” Zhang says. But, he adds, the human resources manager he worked with wanted to examine other factors, such as age, income, marital status, and more. Zhang crunched the numbers on such characteristics, but he was haunted by an offhand question the human resources manager also tossed out: “Is it possible that an employee’s usage of external online communities might influence the employee’s decision to switch jobs?” Zhang’s studies show it is.

Meeting of minds

Online communities are a go-to source for information among many IT workers. A Forbes article Zhang noted in a write-up of his investigation quotes Paul Sammons, a senior SAP business analyst for Baylor, as saying he consults the SAP Community Network before opening a support ticket.

“If you have a problem, the SAP Community Network is a great place to visit and see what other people are doing about it,” Sammons says. “It's also great for day-to-day tips setting up configurations.”

Here’s another thing these communities are good for: determining what answer is likely the best and whom you should trust. That’s because many of these online knowledge-sharing platforms have rewards systems built in that show you how a participant is viewed by other community members.

At stackoverflow.com, a site Zhang says could attract some 90 percent of developers, participants earn points for answers as well as questions. So, for instance, a knowledge seeker may earn a little mark that accompanies his profile to show that he’s asked a question rated “good,” meaning that it earned a score of 25 or more positive votes, or the questioner could be “Socratic,” which means he’s asked a well-received question on 100 separate days. There are even points for poor questions. A “tumbleweed” badge means someone asked a question with a zero score, plus no answers, and few views for the week.

Likewise, there are badges for answering questions and points to be earned by that effort. Like stackoverflow.com, the SAP Community Network also rewards participants with points and badges that signify everything from “I was here” for newbies to “I was helpful,” “I was popular,” and “Super Answer Hero.”

Anyone can see how many points and badges a participant has. Zhang culled that data, then correlated it with job titles from LinkedIn profiles to see which community participants had moved up or moved out after being a part of the online exchange. Not only was community participation correlated with a job change, it also was clear that the type of participation was related to the type of job change that followed.

Role play and work

Zhang explains that there are two types of online community activity he studied: asking questions and providing answers. According to Zhang, employees who pose the questions absorb a lot of learning from the online communities, which contributes to more productivity and better job performance. That, in turn, may lead to higher satisfaction and greater opportunity within the firm. “These are the employees who are more likely to stay with their employer,” he says.

For employees who answer the questions, the opposite is true. Zhang found a statistical increase in the probability that contributors will jump ship. “By contributing knowledge to these communities, individuals may signal their expertise that is otherwise difficult to observe, build their personal brand, and gain outside visibility, which can lead to greater access to career advancement opportunities from employers outside their current companies,” Zhang writes in the paper covering this research.

So, is there a causal relationship between answering questions online and moving to a different job? As a robustness check of his research, Zhang examined this by adding people’s medals and online awards to his statistical analysis models. Going back to the idea that answering questions signals proficiency, Zhang explains that, “If this is the main driver of results, then we should be able to see the medal winners’ participation has a larger impact on job-hopping, which is what we did find.”

Similarly, Zhang looked at the online learners, those who asked a lot of questions. He explains that such curiosity might make people think questioners have little knowledge and, taking that a step further, that they cannot find jobs outside, so they’re staying with their current employers.

The statistical analysis didn’t prove that notion true. Instead, those who asked a lot of questions tended to move up in their organizations. “The people who learn a lot from the external knowledge community improve their productivity and that makes them easily promotable within their current companies.”

Given the results of his research, Zhang recommends employers who worry about turnover watch online behavior more closely. “From a company’s server logs, the company can get general ideas about what websites employees access without infringing on individual employees’ privacy,” he notes.

Then, the company can act to retain top performers. “A job switch is associated on average with about a 10 percent salary increase,” Zhang says. So, companies may want to target those online stars and give them bigger pay increases or promotions to incent loyalty. After all, the high cost of turnover means tossing a little more money in the pot could be a winning move.

The bottom line

• Online knowledge-sharing communities let tech workers ask questions and hear advice from peers.
• Participation in such communities correlates with job changes.
• Those who ask questions are more likely to move up within their current companies due to the proficiency they gain.
• Those who offer answers that are well-regarded by peers are more likely to find a different job with another employer.