Can mobile sensing tools boost worker productivity?
Photo: @criene via Twenty20

Can mobile sensing tools boost worker productivity?

Using smartphones, fitness bracelets and a custom app, researchers have created a mobile sensing system that they claim distinguishes high performers and low performers in the workplace with 80 percent accuracy. The research was supported by the Intelligence Advanced Research Projects Activity (IARPA) within the U.S. Office of the Director of National Intelligence.

In the new system, a smartphone tracks physical activity, location, phone usage and ambient light. A wearable fitness tracker monitors heart functions, sleep, stress and body measurements, like weight and calorie consumption. Location beacons placed in the home and office provide information on time at work and breaks from the desk. 

The information is processed by cloud-based machine learning algorithms trained to classify workers by performance level.

“Mobile sensing and machine learning might be the key to unlocking the best from every employee,” said Andrew Campbell, a professor of computer science at Dartmouth, in a statement.

The accuracy was tested by monitoring 750 supervisors and non-supervisors at a high-tech company and a management consulting firm over a one-year period. Continuous monitoring using the sensor technology was combined with traditional questionnaires to categorize performance. 

Passive sensors promise to offer a “more objective measure of performance assessment” of workers to the benefit of both employers and employees. Traditional review techniques that require manual effort can be burdensome, but also potentially biased and unreliable.

Employees would gain insights into whether their levels of stress, sleep, phone usage or other non-obvious factors are holding them back from becoming more detailed-oriented and disciplined. Future versions could be tailored to specific jobs, provide workers with insights into their mental states during meetings and offer suggestions for reducing stress each week.

The tool could be used solely by workers to gain a private, objective assessment of their performance.

Employers, however, could face privacy and security challenges as well as potential charges of discrimination for using a tool tied to personal data for business decisions. Researchers believe that with incentives and precautions, mobile sensing tools could benefit both employees and employers. Mr. Campbell told The Washington Post, “I’m hopeful this passive sensing technology will be used to empower the workforce rather than used against them.”

BrainTrust

"This idea is ready for research purposes only, to better understand the nuances of what influences productivity, but it should not be tied to compensation or evaluations."

Susan O'Neal

General Manager, Promo Intel & Insights, Numerator


"And so we continue the tragic slide to dehumanization. Interesting that this “research” was supported by the Director of National Intelligence."

Ian Percy

President, The Ian Percy Corporation


"I think this will be super powerful, alerting management of associates who may need additional training before there’s a realized problem."

Shawn Harris

Board Advisor, Light Line Delivery


Discussion Questions

DISCUSSION QUESTIONS: Do you see employees and employers both benefiting from mobile sensing technology that offers insights into worker productivity? Where do you see the obstacles to acceptance?

Poll

19 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
Rob Gallo
Rob Gallo
4 years ago

Regardless of the intent, this will be interpreted as a Big Brother effort. Privacy concerns will be a major issue. In addition, 80 percent accuracy leaves a big gap that can only fuel doubt and lead to a measurable amount of incorrect conclusions and/or bad decisions.

Georganne Bender
Noble Member
4 years ago

“I’m hopeful this passive sensing technology will be used to empower the workforce rather than used against them.” Don’t hold your breath.

Smartphones make it hard enough to get away from work. This adds another crazy layer.

Ricardo Belmar
Active Member
Reply to  Georganne Bender
4 years ago

100% agree — when was the last time an “intrusive” technology capability like this was used for good?

Ray Riley
Member
4 years ago

As the saying goes, “every technology is utilized before it is fully understood.” – and that seems to be much the case with this kind of tech. If the data was solely provided to employees with comparisons and insights, there could be benefits to improving health and therefore productivity. However employers will want a greater benefit, and even with anonymized data there’s privacy risks particularly in smaller teams.

David Weinand
Active Member
4 years ago

No thanks. This will be difficult to achieve any widespread adoption as it just comes across as a way to control workers’ every activity. This would have to reach well into the mid to high 90th percentile accuracy range for employers to consider but that still doesn’t factor in the privacy concerns employees will have.

Susan O'Neal
Member
4 years ago

Putting the creepiness factor aside, I can’t figure out what the purpose of this is – to encourage employees to focus on their well being (thereby improving productivity), or to discourage what the employer believes are unproductive attributes (like not spending enough time at a desk). Are behaviors consistently predictive of productivity? Are the same behaviors equally predictive for all people or do some of us do better with some alone/thinking time vs. others who need to be constantly “on the go” to be productive? This idea is ready for research purposes only, to better understand the nuances of what influences productivity for other people, but it should not be tied to compensation or evaluations.

Ken Wyker
Member
4 years ago

Great idea for research. Bad idea for employee tracking.

Ryan Mathews
Trusted Member
4 years ago

So, this strikes me as “hard science” backed by soft assumptions. Back in the bad old Industrial Age, industrial engineers and social psychologists developed a theory known as the Hawthorne Effect that argued that the process of measuring something changed that which was being measured. So, lets see … you work for me, and you know I’m monitoring every physical function I can, and one of those is stress, and if you have a high stress level you might not get that promotion, and … my bet is your stress levels are going to skyrocket. The other obvious flaw with this theory is that people really aren’t created all that equal. Some, like many professional athletes, perform better under stress, while for others stress causes them to shut down. Ditto for sleep, etc. I’ll think about this again when the assumptions are as “hard” as the measurements. Until then — it’s just creepy technology.

Steve Montgomery
Steve Montgomery
Member
4 years ago

Big Brother has arrived. I know of no one who would voluntarily submit to this type of monitoring nor would I ask them to.

Lee Kent
Lee Kent
Member
4 years ago

In a word, No! Based on this article, I would say this is not ready for prime time. What is one job’s “productive” may be another job’s “not productive.” How in the world would you account for all the variances? It will be interesting to see a few specific applications but, until then, this jury and my 2 cents is out.

Ian Percy
Member
4 years ago

And so we continue the tragic slide to dehumanization. Interesting that this “research” was supported by the Director of National Intelligence. Hasn’t it dawned on these people that there is no “national intelligence?” You can make people perform, that’s how the pyramids were built. But if what you are looking for is commitment, engagement, innovation, perseverance, etc. from your employees, that’s a matter of free will and comes from one’s heart and spirit. No tracking mechanism can measure that. Those qualities can only be given by an employee to an employer, they cannot be taken or demanded. I’m still shaking my head at this news.

Shawn Harris
Member
4 years ago

The folks at Dartmouth have been working on this for at least five years, before many were aware of the impact Machine Learning could have. You can see the original study here. I think this will be super powerful, alerting management of associates who may need additional training before there’s a realized problem.

Doug Garnett
Active Member
4 years ago

This scares me more than anyone can imagine. For two reasons.

The most important thing: If any company believes this system will separate great employees from mediocre ones, that company will deserve the poor results they get. Cathy O’Neil has written extensively about how systems like this are riddled with prejudices — the same ones that bosses bring which get in their way of accurate evaluation.

For example, a Japanese researcher recently investigated the “10,000 steps” idea being used to track employees on healthcare. Turns out that the idea is entirely fictitious — 10,000 was invented by a Japanese pedometer maker who like the way the character for “10,000 steps” looked like a walker. The accurate number is down around 4,000 or so. But Big Brother health plans and companies are already sold on 10,000 — in error.

The second scare is — Big Brother. Nothing demotivates employees more than knowing someone is looking over their shoulder every minute. Please, let’s regain sanity.

Ian Percy
Member
Reply to  Doug Garnett
4 years ago

Your story about the 10,000 steps reminds me of the stories behind the Myers Briggs “type indicator” test or the origins of the familiar eye chart for visual acuity – neither of which had much, if any, validation and yet have been adopted with little thought about their integrity. We are duped from all directions it seems!

Doug Garnett
Active Member
Reply to  Ian Percy
4 years ago

Many of my corporate strategy friends equate Myers-Briggs with astrology for the corporation. Some interest — but nothing that can be applied with the rigidity that many companies attempt. Thanks!

Ricardo Belmar
Active Member
Reply to  Doug Garnett
4 years ago

Doug, you could have led with reason #2 and just stopped right there. I have zero confidence most businesses could be trusted to use any such information like this appropriately — and 80% accuracy? Sounds like a big liability to me.

Ralph Jacobson
Member
4 years ago

Although information gained from this technology can be useful for the employee, I believe as long as the employer has accurate expectations for work output there is little need for this technology for the employer. If the employee is getting their expected amount of work done on time and with a high quality, then they are a good employee from the employer‘s perspective.

Patricia Vekich Waldron
Active Member
4 years ago

This seems like a research initiative at best for a myriad of reasons, including privacy, accuracy and misinterpretation of data.

Ricardo Belmar
Active Member
4 years ago

With only 80% accuracy, this should get a great big “no, thank you!” from corporations. It’s got liability lawsuits written all over it with too much temptation for abuse and bias to be trusted by employees. When technology like this is used to attempt to find those perceived as “no good” or “less than optimal” it almost always fails by misidentifying those that truly outperform as one of those outliers. It only takes one misidentification of a truly great, productive employee to ruin a business. Why take the chance? There are better, less intrusive options.