- 28 Oct 2020
- Managing the Future of Work
Glint-LinkedIn: Worker sentiment informs management
Bill Kerr: It’s late October 2020, and the coronavirus pandemic shows little sign of abating. As governments and businesses contemplate yet another lockdown, the new normal of remote work is taking its toll. Zoom fatigue, isolation, work-life tensions, and competing priorities threaten organizations’ cohesion and productivity. This, along with the Covid recession and social strife, puts a premium on management tools that can provide insights into the worker’s frame of mind. Welcome to theManaging the Future of Work podcast from Harvard Business School. I’m your host, Bill Kerr. Glint, which specializes in employee surveys and AI-based analytics, was acquired by LinkedIn in 2018 as part of LinkedIn’s effort to expand its services. Major firms in a variety of industries rely on Glint’s analytical tools and management dashboards to gauge employee sentiment and to weigh strategy and management options. Can natural-language processing and predictive analytics help organizations pull through this uniquely challenging period? I’m joined today by Justin Black, head of people science at LinkedIn, for a discussion of AI-enabled HR, internal opinion polling, and how companies are tuning in to employee sentiment to guide their strategy. Welcome, Justin.
Justin Black: Thanks, Bill. It’s great to talk to you.
Kerr: Justin, why don’t we begin with a little bit of your background and how you came to join a Glint/LinkedIn?
Black: My background, in terms of education, is in industrial and organizational psychology, or as the tagline says, “The science of making work smarter,” so we’ll go with that. For the past 20 years, I’ve been an analyst, a consultant, a coach in the employee engagement industry. And for the last five of that, I’ve been leading what we call the Glint People Science team, which as you mentioned is now part of LinkedIn.
Kerr: Well, the HR analytics space, which we’re going to be spending a lot of time on—even in the AI machine-learning-based variety—is a pretty crowded field. So tell us what sets you guys apart, and give us a couple of sample customers.
Black: The first one would be the people at Glint. It’s a really people-centric culture, a very well-led organization. It shows in how we innovate, it shows in how we serve customers. And our mission, which is really clear, is to help people be happier and more successful at work. And the second piece is in the execution of that, the realization of that, which is how we bring together people science, technology, and really beautiful, effective user experience to not just make insights easier but to make habit change happen.
Kerr: Okay. Give us a little bit more, as we think about those dashboards and those applications. What are some of the use cases that are most popular with your clients, and how do your customers measure the ROI, or what’s the return that they’re achieving in the use cases?
Black: Our customers are generally measuring success to the extent that we can obviously help them achieve their business goals. So by using Glint, can we improve customer satisfaction? Can we increase productivity in mining organizations, meaning effective time use of equipment? Can we increase the number of bags that get properly handled in an airport? And can we do that by working backwards—from, we need to achieve those objectives through people, and by meeting the needs of people and organizations and identifying what those needs are and meeting them in real time. And so, for some real-world examples, large organizations like Walgreens Boots Alliance—hundreds of thousands of employees across the globe, everything from manufacturing and distribution to retail and M&A. And so they’re using engagement programs to collect regular feedback on the fundamentals of employee engagement. Do people have the resources they need? Do they have effective management practices? Do they get the feedback they need to be successful? Do they feel a sense of belonging at work? They’re using life-cycle programs to understand that people have what they need in onboarding, and if they’re leaving as ambassadors or detractors. They’re using 360 programs to connect the dots to development for certain populations. And so this is a range of employee experience and employee life-cycle, employee engagement, and feedback solutions.
Kerr: I liked how you said your customers, and what you begin with is what matters most for their business success. Does that entail, then, a lot of bespoke customization that needs to happen around what you’re doing with the sentiments and the employee tracking and so forth, to the ultimate outcomes that they’re achieving? Or do you find that there’s a lot more commonality, even if it’s as different as the grocery packing versus machinery utilization versus other predictors of business success?
Black: We have to balance bespoke, customized analyses for customers because relevance is really important. What makes one company successful is not something that makes another company successful. But we also have to help customers scale this. As business conditions change, as our strategy changes, do we need to reinvent an entirely new system to understand if we’re going to be successful? The answer is no. I mean, there are some fundamental truths, right? There are some common outcomes we all care about: customers, efficiency, the health and well-being of our people. And there are common things that predict those that we can track. What’s the strongest driver for one organization or for one team in one organization is not the same as another. But the set of things that we need to track is fairly common across companies. There are some special situations—and this is why we still use consultants in some cases, to look at certain situations and advise on a custom approach. But for the most part, we’re trying to enable our customers to do this on their own so they don’t have to rely on us when they need to pivot their strategy to keep getting insights for what’s driving business success. Is it recognition? Is it resources? Is it the frequency of feedback? Is it the quality of feedback?
Kerr: Let me also imagine potential use cases. One is, I’m a company where our cultures and the employee engagement is going in the direction that we want it. We want to find ways of being ever better, and how you’re engaging in closing any gaps there. There could be another application, which is, I don’t really like our company feel right now. Or the industry conditions have changed such that we need to do a big reboot. And it’s a lot about a transformation agenda. Can your services in Glint fit into both types of a framework—one where you’re trying to continually improve toward an end goal, and another where you’re trying to do something different than what has come before?
Black: Yeah. And I think what happened around the March 2020 timeframe, in terms of how organizations were collecting employee feedback, is a perfect example of this. Of our 500, 600 customers today, the average number of surveys they would send per year on employee engagement, the number of check-ins holistically on employee engagement, was about three. In 2020, it’s been seven, and most of that’s driven by what happened in March 2020, which was Covid. And customers took their regular check-ins on employee engagement, which they do because they never know, we know, what they might learn, and so it’s always good to ask. And they complemented those with some deep dives—five questions on your well-being, your resources, your situation at home—to understand how we can respond right now to needs that you have. And we had some customers do that in some really impactful and quick ways within days of getting feedback, doing things like adjusting flex-work policies, seeing patterns in the narrative intelligence that suggested burnout was going to increase and giving the whole company a day off. And those are much more responsive, immediate activities than these ongoing check-ins that we do in the same way that we regularly check in on our health.
Kerr: To what degree did you find, again, common themes that organizations were facing as Covid set in, and then as we had protracted amounts of time in lockdown. Some of us in many countries, as we listen to this podcast, we’re probably getting ready for a fall wave that could be as bad as the one in late spring. How much of that was consistent with the clients that you were experiencing—it was something that was just inherent in the nature of working from home and remote and being a little bit more isolated—versus industry-type specific challenges?
Black: I think where we saw differences was in job types. There were certain types of workers who were still going to work. And we would hear the comment early on in March 2020 about returning to work—“When are we going to return to work?”—sort of late March 2020—“When are we going to return to work?” And all of these people who were still working … And what we meant was return to what we thought was a normal way of working. What we saw early on in the data was pretty expected. Our system tracks 144 topics, looking across all of the unstructured comments and feedback that employees give in the system. So if I say, “I’m so exhausted that I can’t get my work done,” the system sees that and tags it with burnout. And we can look at the rates of those tags. And so the rolling average on burnout and how you look at it is anywhere between 2.5 percent and 4 percent. And the rate leading into March 2020 was 2.4 percent. After Covid, in April 2020—after the initial big reduction in or change in how we worked—the rate went up to above 5 percent, so more than double. It has since increased steadily. And so one thing that we’re able to see by tracking—we’re now tracking, I think, 5 million responses since March 2020, and it’s now October 2020—is how these things change over time. And so, while some things have remained the same a couple of things have changed. Burnout is one that has increased, and that doesn’t bode well for the fall of this year. If Covid, as expected, continues to cause more reduction in privileges and people have to stay home and they can’t go into the workplace and businesses continue to tighten, these burnout rates could have even more impacts on productivity, on well-being, than they had the first time around. And so Covid fatigue is not just, “I’m tired of Zoom,” this may be fatigue that turns into actual burnout, which is a significant issue. One of the surprising things we saw—and this was good news—is that people’s ratings of collaboration actually went up. So there was this rallying effect of, “We’re doing this a new way, we’re all in it together, we’re going to figure this out.” And people actually figured out how to collaborate really effectively online. And it hasn’t been universally true, not at every company, but on average, that’s remained really good. Another really important driver—so, if we look at not just what are people saying but what causes people to say that they are going to be burned out and what causes people to say that they are going to be resilient?—one of the most important factors is balance. And although workload is important, balance is almost twice as important as a driver of burnout. So how do we use that information to give more flex to employees? That impact hasn’t changed, and the scores haven’t changed much over time. We are making changes, but people are continuing to have needs come up. The point being: We have to keep asking them what they need and responding in time. It’s not enough to just ask and move on. And one more big theme: Employees who feel their managers are ineffective communicators are 2.7 times more likely to use language that signals burnout. We say to managers, “Communicate until your team tells you you’ve over-communicated.” And it’s so rare for that to happen, and it’s really important for that to continue.
Kerr: You talked about how the forward-leaning companies were surveying their employees about three times a year. For 2020, we’re already double that. And that can raise, first, the question of how much is the right amount? You can be too little, you can be too much. How does that get calibrated? And the second is the potential concerns of employees about confidentiality and the protection of their data.
Black: I’m so glad you raised this issue, because it’s central to the success of any feedback program, is that we have a system that people trust and that people believe helps them be happier and more successful at work, and not just create more work for them, right? It’s not just a check-the-box compliance exercise, or even worse, something that I don’t trust. I mentioned we’re seeing more surveys deployed than ever. So I’ll tell you that, if employees were to give leaders feedback that they’re being over surveyed, leaders would stop surveying. It wasn’t easy for us in the early days to get leaders to survey more frequently. We had to show them that there was a way to do that, that employees actually liked and wanted it. And I think a lot of the evidence around Covid would suggest that, if you do it the right way with the right intent, the act of asking actually can be almost as important as what happens next. And I say that because a lot of the feedback we got from employees, or leaders got from employees, during their initial pulses on how employees were feeling in March– April 2020, right after Covid was, “Thank you so much for asking. It just means so much that, even though you know the answer is going to be, ‘We’re not doing so hot,’ you actually asked us and asked us what you might be able to do to help even a little.” There’s some evidence that companies that are doing this regularly have been able to do it effectively and make it a constructive process. How do they do that? Part of it goes back to just hard science, right? How do you design a system that maximizes signal, reduces friction for the employee experience—both from a user experience perspective—How do you lay out the question? What are the colors on the page?—but also from a measurement model perspective—What’s the smallest number of questions we can ask to get the maximum signal? We really value employees’ time. As a nerd and somebody who loves models and prediction, I would love to ask employees 70 questions every week, measuring various aspects of their employee experience, right? As a leader, I would like to see maybe one data point a day. And so how do we connect the dots between those two things? The second piece of this, obviously, is data privacy. And so making sure that how we present data back, store data, transfer data, use data, is completely consistent with the General Data Protection [Regulation] requirements. And even beyond that with expectations that people have for how to just handle their data in a good way is the other more boring and most important foundational piece of that.
Kerr: Continuing on the data side—one nerd with another nerd, possibly, on this podcast—you could think about not only assessing where a company is in an absolute sense, but also where should they be? And also, what are the things, the root causes, that are getting in the way? How do you guys go into that diagnostics? And how much do you need in order to be able to start identifying the gaps of where a company really should be aspiring to be, and what are the steps that would help them approach that?
Black: The question about how do you combine the various pieces of information we might have to serve up recommendations to leaders is one we wrestle with a lot, because on the one hand, we might want to be highly prescriptive and use a lot of data; on the other hand, doing so assumes that we understand the impact of each of those pieces of data and the outcome each leader is trying to achieve. And that’s a pretty big assumption. And so what we try to do is use things we know are important—and you mentioned a lot of them. And those are a lot of the things that go into the formulas that serve up recommendations for managers. For example, if I’m looking at an employee engagement survey, and I’m a manager, my team has just given me feedback, the system’s going to look at, for my team, which of the pieces of feedback, which of the items that we measure, has the strongest impact on the outcome. Usually it’s some overall measure of employee engagement. Which of them is highest and lowest compared to some sort of benchmark?—whether it’s the company average, which is actually the best benchmark, it’s the most fair to everyone—or an external benchmark—How are things comparing over time? And you want to add all that up in a fancy equation and spit out a recommendation of both things we want to maintain and things we might want to work on. But we don’t go so far as to say, “This is the one thing, and it’s definitely this.” We say, “The data suggests it’s probably one of these three things.” And we give leaders a dashboard and basically a PowerPoint export that they can take to their teams and have that conversation. “The data suggests it’s one of these three things. What do you guys think? In your feedback, you said this; what did you mean by that?” And we give them those questions to ask so that it becomes not just a data-driven process but a data-informed habit-change process that involves the team.
Kerr: The data’s used for, to your earlier point, the communication that the leaders need to do, but then also as a way to spark that follow-on conversation as to what’s what lies behind this, the root causes to whatever you’re observing.
Black: Yeah, exactly. And to the extent that we can get that prediction down to the most local team, the better. And so that’s where data-processing speeds and more efficient formulas have really helped into localizing this information—that used to only be available at the company level, and then to cascade it down as if it applies equally to everyone.
Kerr: You mentioned earlier “burnout.” And if I had to have you think about the many different clients that you’ve worked with, what are the biggest things in the post-Covid world that companies are struggling with? And what’s maybe been something that their team really is struggling with but maybe is less visible to the leader of the team or the organization?
Black: There are some key things organizations are struggling with that I think we can learn from that are universal. You don’t necessarily have to check on this; we might just assume it’s true and go work on it. One of the things is belonging, is the people’s sense of belonging. One’s sense of belonging has always been strongly correlated to people’s ratings of their happiness and success at work. That impact for the first time I’ve ever seen in 20 years in this industry is now stronger than the impact of how I feel about my career-growth potential at this company, which has previously been one of the strongest items—predicted engagement. And so people feel involved right now is really key, and not everyone’s scoring high on that. And certain populations are scoring much lower on that—Blacks and African Americans at senior leader levels, women at senior leader levels, right? You have hotspots in organizations that need to be surfaced and addressed in an authentic way. And that’s a journey a lot of companies are on right now, and it starts with knowing where we are. Another huge theme that we’re seeing—and it’s something that I think people are struggling with—is this balance of support and empowerment and micromanagement. When it’s working, employees feel like they have flexibility, they have clear goals, as long as they’re getting their work done they’re okay, if they need help they can get it. When it’s not working for companies who are struggling, we’re hearing people say, “I want my manager to leave me alone; it’s just too much. Trust me, I’m getting my work done.” Or, “I’m not, because I have four kids at home, and we’re trying to do school, so I’ll get it done later.” And so that’s a really big theme we’re seeing, is around trust and empowerment. I’d say the other big one, besides burnout, which we’ve already talked about, is connection. Employees who say that their employer is helping them feel more connected are four times more likely to feel well supported generally, which we know is critical for mental health. And so employees do expect, there is this underlying expectation, that the company is going to help me feel more connected. And so I think the other thing we can learn is not to underestimate how much employees want companies to help them feel connected, as we go into a continued—and maybe even increasingly tough—work-from-home or work-from-wherever, or try to work out in the real-world environment in the fall of 2020.
Kerr: Those are consistent with some of the managerial themes that we’ve heard on this dispatch series of leaders—I think it’s pretty straightforward in this April timeframe—of needing to step up and create this glue and rebuild some processes for the short run. But as the summer has kind of worn on, and as we enter the fall, and as we think about some of the challenges they have until, ideally, the vaccine comes and we can start to resume a normal life, leaders are going to have to step up evermore. And you’re highlighting some of those belonging themes that are necessary for us to hold that glue together. Let’s dwell upon something that you highlight in describing the belonging. You mentioned that the Black senior leader and also women, and some of the questions that they were raising about the belonging. Talk to us a little bit about the ways Glint’s customers use your tools for diversity and inclusion efforts, which should always be at the front burner, but especially right now are being highlighted as very important areas for businesses.
Black: It’s such a hot topic right now, and I’m so glad. It’s something that I did my first research project on, I think, 15 years ago. And I don’t think interest in it increased much between then and about eight months ago. Sometimes crises like Covid present this opportunity to bring attention to an issue that we know is important. The good news is that any good employee engagement survey will have some measures of inclusion and involvement and hopefully fairness or belonging as part of it. And so our customers have been checking in on a regular basis on whether or not their employees feel involved and valued and heard. Most platforms today, including ours, will surface up if you have hotspots or bright spots, with regard to those questions and certain attributes or certain populations. I think where it’s changed is that the number … in a couple of ways. For companies who are already checking in on diversity and inclusion and belonging, they’re asking more questions about it. For companies who weren’t checking in on it, they’re doing deep dives. “Maybe I don’t have an engagement survey, but this issue is important. I’m going to start by just asking seven to 10 questions just on this topic.” But what everybody is doing, and what we’re seeing increasingly, is just serving up, coming off of the survey, cascaded recommendations for what I can do as a leader, three simple things, four simple things. For example, “I see your scores on belonging are fairly low. Here are three or four simple things that you can do in the next couple of weeks to get started on the conversation.”
Kerr: Let me take us to another part of the workforce and something that we’ve been thinking a lot about at Managing the Future of Work, and also we see a lot more activity today, is around blended workforces that have full-time, part-time, contract workers, especially around projects building up. To what degree do you use your metrics and your surveys to touch on all different parts of the way a worker may now interact with an organization? And is it important for a company to think about something like “belonging” when they’re speaking about a gig worker that is with them for maybe a few months, maybe longer, but is not the traditional full-time employee?
Black: There are some simple things when we’re talking about different populations within the organization, and then there are some more complex things. And the more complex things tend to do with contractual agreements between the employee and the company that we want to make sure we mind. On the simple level, if we work backwards from, there are key populations that we know are different in our organization for one reason or another—they’re part time, they’re remote versus not, or what we call at LinkedIn “home employees” versus “office employees”—and we track those things. And that’s simple enough. It gets a little more complex when the contractual arrangement between the employee and the company is, say, through a third party, in which case you have to make sure we actually have the right to ask them these types of questions. And so our customers try to involve everybody unless there is just some reason they can’t. And if they can’t involve people, it’s usually for a good reason. There are good laws protecting contractors from doing work that isn’t related to the contract—albeit, generally, our employers, our customers, are trying to involve everybody. And the good news is, the questions we’re asking are fundamental to the employee experience, right? “Do I feel a sense of fit and belonging here? Am I in the right job? Do I have what I need to be successful? Do I do work that’s meaningful to me? Am I learning and growing?” These things are important to everybody. The important thing is to make sure you know upfront what “depends” questions you want to ask on the end, so the feeling that people have about recognition here “depends” on where they work or what kind of an employee they are so that you can slice the data on the backend and track those populations.
Kerr: It’s a complexity many organizations will need to be better able to handle toward the future. Are there any particular trends in human capital management and the ways that technology is connecting into the workforce that are top of mind for you as we go into the next year?
Black: The thing that’s occurred to me lately is that, over the past 20 years, I’ve seen a lot of crises, a lot of shifts in how we work, how we approach feedback, what technologies we use. What seems to remain true through all of that is that companies who take the time, leaders who take the time, to convert their intentions into culture through habit change and actually do that hard work are the ones that tend to be successful regardless and use the new technologies in the best way to serve the needs of their employees. Those habits are starting to be pretty consistent over time. The most important one I see is conversations. What’s the expectation we have in this organization for when we’re going to check in, what we’re going to talk about, and who’s going to be involved in that so that we can learn and grow and have what we need to be successful? The second one is feedback. Are we a feedback-seeking culture, first and foremost? The third one is agile goals. Do we set achievable goals? Are we willing to change them when conditions change without sweating it too much? Do we communicate about them regularly? And the fourth one is changing through learning—has learning new things been open to changing what we’re doing? These four buckets of habits seem to be pretty consistent across organizations that thrive, even in tough times.
Kerr: Justin Black is the head of people science at Glint/LinkedIn. Justin, thanks so much for joining us today.
Black: Oh, Bill, I had so much fun. Thanks so much for inviting me in.
Kerr: We hope you enjoy the Managing the Future of Work podcast. If you haven’t already, please subscribe and rate the show wherever you get your podcasts. You can find out more about the Managing the Future of Work Project at our website, hbs.edu/managing-the-future-of-work/. While you’re there, sign up for our newsletter.