Managing with Technology

Anyone who manages even a small number of workers could probably describe the complexities involved in balancing the needs of their employer, and those whose work makes the company go. Each person has interests in personalized continuing education and skills development, clear channels of communication with colleagues and supervisors, and work-life balance issues like flexibility in scheduling.

So, how to turn these many variables into the kind of insight that not only helps create a mutually advantageous workplace about even one person, let alone scale that to hundreds or even thousands of people?

"Data is noisy, and data about people is probably the noisiest," said Tom Walsh, senior director of data science at Kronos, a Lowell, MA-based workforce management technology company. "We will never have perfect insight into why they made certain decisions or took certain actions, and we have to admit that upfront. But there are a lot of areas where collecting data about what people are doing, and using that to help streamline their processes, can be really beneficial."

Can AI Make Sense Of It All?

For Walsh, harnessing the power of machine learning and artificial intelligence is a new given, in trying to improve a company's business processes and an employee's own experience.

"There has been a real sea change in both the technical side, we could talk about the revolutions in deep learning in the past few years, but also a sea change on the business side in terms of acceptance of these applications from both management and employees," Walsh said.

Implementing predictive analytics for common business processes is nothing new; nor is the periodic gauging of employee sentiment through vehicles such as surveys. However, HR consultant Josh Bersin thinks HR engagement technology has entered an entirely new era in recent months with the introduction of "nudges" to people throughout an entire organization, not just HR executives or managers. Bersin calls this new paradigm "Engagement 3.0."

Bersin cites Mountain View, CA-based Humu as the exemplar of the "3.0" model.

"The origins of engagement surveys were top down: managers asked employees what they thought, and as a result, the managers were supposed to act," Bersin wrote in his blog. "Well, now we move to engagement tools for everyone. This is what I think of as Engagement 3.0—using feedback and data to give everyone tips to perform better.

"In the case of Humu, the product is based on a 'nudge engine,' which is a set of carefully developed rules, hints, tips, and suggestions that help people change behavior in small, meaningful ways."

Humu has, in fact, trademarked the Nudge Engine, and cites greatly increased employee engagement when management not only uses the technology, but shares the fact with employees: "When managers share Humu's findings with their teams, workers have rated them as 46% more committed to change—and also registered the company itself as 22% more committed – than those whose managers didn't discuss findings," Humu CEO Laszlo Bock wrote in introducing the Nudge Engine. "Surprising? It's science: transparency in leadership is directly linked with employee commitment, effectiveness, and trust in leadership."

While nudges related to aligning individuals' goal with corporate objectives may be new, they have been used in benefits administration for some time, using the principles of behavioral economics; these postulate that a person's unhealthy present bias can be overcome with persistent healthy nudges. In fact, one of Humu's co-founders, Jessica Wisdom, practiced healthy nudges while at Google, devising ways to display snacks in such a way that healthier options were more visible. Now Wisdom, Bock (Google's former chief human resources officer), and Humu's third co-founder, Wayne Crosby, are betting nudges, based in good part on the work of Nobel Prize-winning economist Richard Thaler, can transcend individual goals to the Goldilocks balance of creating synergy between corporate goals and employees' personal happiness.

"Humu's primary measure underlying the model is employee happiness, which is shown in a rich trove of academic research to be a driver of outcomes like performance, retention, commitment, and so much more," Wisdom said in an email. "We take into consideration as many signals as each of our partners decides to provide, but—yes—these can include things like tenure, job type, and performance data, or be as granular as commute times."

The Humu platform then combines this information with the sentiments collected in the Humu employee diagnostic —a short set of statements about the employee's experience of work that can be answered on a seven-point scale of "Strongly agree" to "Strongly disagree," plus a few open-ended questions (Humu recommends the diagnostic be administered every six months).

That combination of data, Wisdom said, allows Humu's analysts to "better understand, and control for, the variables that are at play. What our system identifies as most important to teams is unique to each of the organizations we work with."

One early Humu customer, Culver City, CA-based restaurant chain Sweetgreen, discovered its frontline employees wanted more career development opportunities than they felt they had.

"After the assessment from Humu, we found that while employees were pretty happy, career advancement was a spot we need to work on," Sweetgreen co-founder Nathaniel Ru told First Round Review. "The software sends 'nudges' to remind head coaches of more specific topics to address in this realm, which could be covered in a Sweet Talk or by holding one-on-one conversations with the restaurant team members."

Will it scale? Whose interest does it serve?

While more vendors introduce AI-enabled HR technology—Kronos and IBM recently announced a partnership that will combine IBM's Watson Career Coach and Kronos Workforce Dimensions platforms to help businesses engage hourly employees on topics such as training, shift scheduling, and career coaching—its proponents must find a way to balance the promise of creating happier individuals with where the real payoff comes, in scalable recommendations that don't slip into manipulation.

"I think it helps both the good manager and maybe not so good manager, but you have to remember all this is happening on the margins," said Todd Haugh, assistant professor of business law and ethics at the University of Indiana. "You don't need to make a mediocre manager a great manager through a nudge. Nudging one person doesn't really matter. Machine learning with 1 million inputs affecting 10,000 or 20,000 or 30,000 people or more, that's when you see real benefits and also why this technology and tying it to behavioral science is so intoxicating. They can see at scale this stuff means real dollars and real benefits to companies, which also introduces a real risk for unethical action."

Humu's Wisdom said the company is prepared to demonstrate its technology can provide organization-wide insights and benefits, and still protect the individual. "While we have worked with organizations that range in size from several hundred to over 50,000, our system works at its best with large enterprises," she said. "Data analysis identifies patterns in data; more data makes it more possible to find meaningful patterns. We have strict agreements with each of our partners about how their employees' information is used."

According to Humu head of security Bryan Zimmer, the company has built the platform to comply with information security specifications published by both the ISO and also the American Institute of Certified Public Accountants. In a recent blog post, Zimmer also said the platform was built with the European Union's General Data Protection Regulation (GDPR) in mind "from day one, and we extend the rights from the GDPR to every user no matter where in the world they live. This includes transparency, control, and data portability."

Though Haugh has some concerns about the potential for nudges to be misused, he is also cautiously optimistic about the upside.

"I'm really not a 'no' guy on this technology, and certainly not the use of behavioral science within companies," he said. "It can be an incredible, important tool. It's cost effective and has the opportunity to positively impact the lives of lots of people. The idea of nudging a manager to consider the well-being of their direct reports or their hourly employees is a fantastic idea, and as the technology improves along with machine learning, we'll be able to do that better and better, and that is incredibly positive."

Gregory Goth is an Oakville, CT-based writer who specializes in science and technology.