Fueling People Analytics with Feeling

Now available: People Analytics For Dummies

“When I say it’s you I like, I’m talking about that part of you that knows that life is far more than anything you can ever see or hear or touch. That deep part of you that allows you to stand for those things without which humankind cannot survive. Love that conquers hate, peace that rises triumphant over war, and justice that proves more powerful than greed.” – Mr. (Fred) Rodgers

In putting together material for the People Analytics for Dummies book recently I have been sorting through a database of survey questions I have accumulated over the years. It strikes me how many of the questions I have found the most mathematically useful (predictive) of important outcomes are about feelings or inherently related to feelings.

A few good examples:

• I am proud to work for (Insert Company).
• I am inspired by the people I work with at (Insert Company).
• The work I'm doing here at (Insert Company) has personal meaning to me.
• I am motivated to do more than expected to help those I work with succeed.
• I can recall a moment in the last three months when I felt happiness at work.
• I feel a strong sense of belonging at (Insert Company).
• I enjoy coming to work every day at (Insert Company).

It is not lost on me how paradoxical the power of feelings is when I consider that the job of people analytics is to help make decisions about people more rational. Presumably making decisions more with data and math should reduce sensitivity to feeling, not increase it, but I have found just the opposite.

This has got me thinking lot about feelings. Consequently I have been seeing feelings everywhere. 

The Mysterious Case of Mr. Rodgers

On Sunday I went to see “Won’t You Be My Neighbor (trailer)”, which is about the philosophy, life and legacy of Fred Rogers, the beloved host of the popular children's TV show "Mister Rogers' Neighborhood”, which started on U.S. public television in 1968 and ran for 895 episodes. It is very good. Bring some tissues.

By all accounts MR. Rogers would not have been picked out of a lineup as someone who would be successful on television, yet somehow he was successful beyond anyone’s imagination. He received the Presidential Medal of Freedom, 40 honorary degrees, and a Peabody Award. He was inducted into the Television Hall of Fame. He was ranked number 35 of the TV Guide's Fifty Greatest TV Stars of All Time. Several buildings and artworks in Pennsylvania are dedicated to his memory, and the Smithsonian Institution displays one of his trademark sweaters as a "Treasure of American History".

If you grew up in the United States anytime in my lifetime you know him well. If you don’t know him, Fred Rodgers philosophy might be summed up by this remark he made once in an interview: “Love is at the root of everything, all learning, all relationships, or the lack of it.”

With all the emotionally disturbing problems in the world today there has been a bit of nostalgia for Mr. Rodgers. Outside of helping children deal with crisis he didn’t venture at all into the realm of politics. However, one of the more most salient articulations of the power of feeling was an impromptu argument he made before Congress. Congress wanted to cut funding. Fred Rodgers was asked by PBS to attend a Congressional hearing to help defend PBS's funding. Fred seemed uncomfortable in this setting. He had prepared a statement but after hearing the congressman make a remark about how he was tired of people reading statements Fred decided not to give his statement. Instead he made some brief impromptu remarks. His argument was this: “If we can make it clear that feelings are mentionable and manageable we will have done a great service for mental health.” Finally, he stared into the eyes of the often-bellicose congressman and closed with the lyrics of a well-chosen song that he had himself written. It seemed to be a perfect conversation between the inner child of Fred Rodgers with the inner child of the bellicose senator. The senator threw up his hands and Fred got PBS their $20 million dollars. Watch the interaction now (5 mins).

The Rebirth of Emotional Labor

I also am reminded of a book I read as an undergraduate in sociology by Arlie Russell Hoschild, called “The Managed Heart: Commercialization of Human Feeling”. I have always believed this book made a profound contribution to research about work. It is one of my favorites. I had imagined it resigned to the dusty shelves of sociological antiquity when I ran across an article that spoke of it yesterday : “How faking your feelings at work can be damaging”. It turns out there is new research building on Arlie’s work and exploring how burnout relates to emotional labor. For some catch up: Emotional labor is “the effort that goes into expressing something we don’t genuinely feel. It can go both ways – expressing positivity we don’t feel or suppressing our negative emotions.” 

The focus of the previous work on emotional labor was customer facing occupations that require the management of emotion, however more recently they have begun to think about how all work that requires human interaction have some potential for emotional labor. We have become more interested in emotion as we have become more interested in something in the professional zeitgeist today we call “burnout”.

It was previously thought that burnout was a result of the volume of overall activity without break – e.g. a work life balance theory. The new theory is: “Those who report regularly having to display emotions at work that conflict with their own feelings are more likely to experience emotional exhaustion” (e.g. burnout). The implications of how you prevent or treat burnout are entirely different depending on whether you believe burnout relates more to the volume of overall activity or the emotional experiences of employees during that activity or how employees manage their emotions during that activity.

There are many questions to be answered. What emotions do people feel at work? How frequently do people feel emotions at work? How strong are the emotions people feel at work? When do people feel strong emotions at work? What causes people to have strong emotions at work? How often do people feel strong emotions? What are the positive implications of strong positive emotions - socially and individually? What are the negative implications of strong negative emotions – socially and individually? How do people’s sensitivity and response to emotion vary? How can companies produce more positive emotions? How can companies reduce negative emotions? How can companies help employees deal with emotional dissonance more constructively to reduce burnout? What is the business case for dealing with this problem? etc… 

The First Paradox. Taking Control Away From Feeling Produces Positive Feelings and Reduces Negative Feelings.

In People Analytics we use data to make decisions that are based primarily in evidence as opposed to decisions based primarily in feeling. We do our work primarily for the benefit of corporations. Doesn’t this seem inhuman and potentially devoid of feelings? Doesn’t this forebode findings that conflict individual needs with the needs of corporations? The paradox is that it produces exactly the opposite.

I have always found that when you use data to make decisions about people in corporations the decisions seem to always end up working out better for the individual people and the corporation. This insight has produced strong feelings in me (profound meaning) which motivates me to do more work, even under difficult circumstances. ’ll spare you the sermon. Yet, it remains a paradox how two things most people believe are opposed: worker needs and corporate profit are in fact not opposed at all. I find this is true broadly and for me in my own work for others.

There is various mechanism through which making decisions about people also reduces negative human feelings. One of the more important mechanisms for this is fairness. Humans (and it turns out other primates as well) are concerned with fairness. Data help the corporation reduce individual biases, which squares up to the group need for fairness, which reduces negative personal feelings. Data can help us make better decisions, provide more certainty about this and provide more transparency. Negative feelings can be reduced when decisions are made in this manner. 

It doesn’t stop here but this is one example.

The Second Paradox. Discarding the Need to Interpret Feelings as Wrong or Right Produces Positive Feelings.

Individual feeling may be wrong or right (or both right and wrong), and at the same time a fact. How is that sentence for the start to a section? A little confusing I’m sure. Let me elucidate.

Here is the problem. While personal feelings can sometimes be taken for the wrong reasons they are never wrong as a statement about the current state of the person. They may seem wrong to another person, or wrong from more objective perspective, but as a descriptor of a subjective state of a person they are always an objective fact. Feelings are both subjective and objective when viewed in this light. The person has a feeling, and this has consequences for them and others – arguing whether the feeling is wrong or right is not useful in most circumstances. It is better to acknowledge the feeling and work with i, than to ignore it or discredit it. 

The paragraph above may just sound like an interpersonal philosophy that has nothing to do with people analytics, but it is also important in people analytics. The misconception is that we should throw out analysis of feeling on the basis of the inherent subjectivity. This is the exact opposite conclusion you should take. If the behavior of individuals hinges on the subjective interpretation of events, and this subjectivity varies, then we HAVE TO understand the subjective variance. The measurement of feeling provides a window into important variance in understanding and predicting human behavior.

To dismiss feeling over concern about subjectivity is a grave mistake for people analytics and companies. If you depend on a person it does not matter if you agree with their feelings when they leave you in a lurch and as a result you lose thousands of dollars. Whether you agree or disagree you still may find it useful to prevent, predict or associate information about their feelings. Ignoring them gets you nowhere. Measuring feelings provides you with an analytical path to understand human behavior, predict human behavior and change human behavior. 

Feelings are important to analyze, regardless of whether you or the executives you support agree with them.

The Third Paradox. The Extraordinary Rationality of Irrationality.

By most accounts we would expect rocket scientists to be the most rational and therefore the least emotional people on earth. Maybe with rocket scientists you just need to make a mathematical argument that you are paying them y output and you expect x inputs in return. How then would you describe this actual footage of rocket scientists? 

The Falcon has landed

What 400 Very Happy Rocket Scientists Look Like

I can only imagine the amount of exceptional labor produced by these group of extremely “emotional” “rational” people. I can only imagine they will reach mars or wherever they want to go. It is only a matter of time. 

How would you like to have a group of employees that work together like this?

Feelings are as powerful as rockets and more. Having worked for several cultish companies I believe that harnessing the motivation of people at work through meaning is how companies that succeed over long periods of time manage to operate so well. 

The Fourth Paradox. Feelings are Chaotic but Valuable Predictors. 

Feelings can compel people to do great things or stop them in their tracks. Feelings can provide us energy or make us sick. Feelings can compel actions to buy or compel actions for relationships to die. Feelings determine the fate of people and organizations. 

How well or poorly companies understand and influence feelings is important. This is true in Marketing and this is true in People.

The Movement to Ignore or Abandon Feeling in Analytics is Paralyzing

In recent years there has been thrashing for some way, any way, to get insights other than a survey. We have seen surveys reduced to a single question. We have seen surveys moved from once a year to every other year. We have seen increasing interest in text mining / natural language processing (NLP). We have seen organizational network analysis (ONA) applied to email meta data and even the content of emails – e.g. continuous scanning of employee conversations to identify signals of relationships and sentiment. We have seen efforts to use voice or video cues to interpret emotion. We have seen increasing interest in machine learning. We have talked a lot about big data, social data and artificial intelligence. We have added data scientists from backgrounds like engineering, statistics and computer programming. 

Many of these new adaptations of the field are useful, but not all of it is universally useful, and none of this implies surveys are not useful as an important channel to understand people. I suspect that this movement away from surveys is based in part on an unfounded interest to find “more objective” data sources than surveys under the view that this will make the field of people analytics work better. Nothing is further from the truth.

Some thought leaders have gone as far as to object to calling the field people analytics at all because they believe that in the future of work will increasingly shift from people to robots*. They have another term be used that doesn’t include people at all: workforce analytics. 

My view on this is clear. It is true that some work is going to be increasingly performed by machines, but this does not require we dissolve our specific interest in people just yet.

People Analytics is uniquely and unapologetically about the aspects of the company that are performed by people. There need be no panic because there are is an increasing amount of work performed by machines - we still have to answer questions about people. At the point at which no people are left I will surrender my argument, but at that point this argument will cease to matter so it is mute. Until the day people cease to matter at all we need some disciplined empirical methods designed specifically for people in organizations. Specifically, people analytics is the intersection of people science, strategy, statistics and systems. 

Frankly, what astounds me about all the thrashing in people analytics that in my 17 post-graduate career working with data of all types across more than 7 employers in different sized companies, geographies and industries I have never encountered a situation where subjective survey measurements did not produce valuable insight for the management team. In my experience surveys are instrumental, if not fundamental, to the most successful companies. In some cases the surveys were responsible for the most important business insights across all sources, not just HR data sources! After all, the most difficult part of organizations to copy are people.

It would be lovely if you could stick in place and assume we operate as robots, Jet Engines, Mechanical Systems, or Websites. Unfortunately, perhaps fortunately, people and organizations don’t work this way! People have individual perspectives, preferences, strengths, weaknesses, motivations, knowledge and emotions that form together into collective cultures and outcomes. While tempting to dismiss emotions the premise is rooted in a lack of awareness in how the human machine operates and the value of surveys in understanding, predicting and learning how to influence people. In this it turns out that feelings are very important.

Try as hard as you may to avoid it the fastest and cleanest path to understand feelings, at least consciously, is to just ask. However, if you want to analyze the information you need to scale and code it. If you need to scale and code it, then is usually more accurate and efficient to get it directly. That is a survey.

If You Want to Use Feelings well in People Analytics, the Details are Important.

I suspect the different results people are getting with surveys is more about the design nuances (why, when, what, and how) of survey research than it is about the value of survey research. A survey should be a design element in broader research agenda, not just a survey for the sake of running survey. While the absence of survey may absolutely exclude certain insights, the presence of survey doesn’t insure certain insight. It is not a talisman.   

I suspect that while some organizations claim their employees have “survey fatigue” and others do surveys all the time and achieve enthusiastic response rates (unnecessarily high response rates) also has something to do with both the quality of the survey and how it is used. My experience has been that to a degree 10x any customers (or any other type of person in the world) employees are motivated to provide companies with feedback. Employees also have lot of feedback and there feedback is constantly changing. If there is no integrity about this, no accountability, no action and no sense about this then of course they may give up trying. If this happens you have already lost. It is like saying we can’t do CPR because he is already unconscious. O.k., Einstein what’s your plan then?

People are attune to our shenanigans: they will feel uncomfortable or reluctant to “uncanny valley” of repeat poorly designed surveys. How you arrive at the questions you use is important. How you evaluate the performance of the questions you use is important. How you adapt and change your questions is important. What you do with the answers to the questions is important.

There is a professional craft and follow through. Just because someone can draw, doesn’t make them a great architect. Nor does having a pencil. If you think the work of surveys is just getting the survey launched and the data collected, and actions planned, again, you are gravely mistaken. You are here, but you are not with us.

For example, an opinion survey should be viewed as a single component in an objective longitudinal human experiment. You need to be able to see before the opinion, the psychological change in opinion and the consequences of the opinion. Another way of saying is that you don’t start when you start the survey and you’re not done when you end the survey. 

If you use your survey as a one-off study of opinions you have limited your ability to associate it to objective precursors and consequences so that you cannot obtain the perspective you need to interpret the opinions you just collected.

This shift requires nuanced differences in workflow and statistics to interpret the information you receive. The secret to success is really in how you go about the data. For example, you cannot analyze longitudinal if have aggregated surveys in a point in time cross section, segmented it and then abandoned the individual data. Longitudinal analysis requires you maintain data at the individual level in a database over time so it can be associated to other data sources and outcomes. You also must be able to apply statistical applications to this data. To be clear, all this MUST be done while applying a careful confidentiality protocol, or work with a third party who does this, so that nobody from the company ever sees the data at the individual level. 

Many leaders are either unaware of the importance of data workflows I am talking about, don’t have teams with the technical capability to apply the data workflows I am talking about, or are unwilling to explore the data workflows I am talking about for fear of something different. Whatever the reasons if you are unwilling to learn and work through these issues you will have resigned the companies you work for to the dust bin of business history. The emotionally paralyzed companies will be systematically dismantled by organizations more adept at working with people. All this will take less than 20 years and if you watch the recent moves of Facebook, Amazon, Google, Netflix, Microsoft, Airbnb, Uber, SpaceX and Tesla it may in fact be less than 10.

The misconception that surveys are not useful has made its way into the zeitgeist of people analytics and must be dealt with. The primary thrust of the anti-survey argument is that surveys are subjective, emotionally squishy, not fact, possibly wrong, and THAT is why some organizations are not getting results with surveys. Paradoxically, nothing can be further from the truth... 

Perhaps my remarks here are too detailed for some or too abstract for others and for this I sincerely apologize. I am doing my best for a blog post – if I am unclear there are other ways we can get on the same page. Regardless, it takes time though, you have to invest time.

Keep an eye out. I plan to have related articles on surveys. I plan to include 10 useful survey questions on ten useful topics in the People Analytics for Dummies book coming out in December. 

I also plan to address in detail the major opportunity, as I see it, with surveys in Talent Attraction and Acquisition. While there is always some good in measuring satisfaction – I believe the major opportunity is understanding the emotions that drive people into the recruiting funnel to enhance recruiter’s ability to tap into them better. More on this in my live online Talent Acquisition Analytics workshop in July.

I am also interested in the emotional experiences of all people involved in the recruiting process (candidates, managers, interview teams, and recruiters) and I’m particularly interested in of inclusion and fairness. 

There is also major opportunity to expand the use surveys in on-boarding – not just as a smile sheet for orientation, but to truly measure important changes in the motivation over the first year. I want to measure when expectations are not set right in recruiting, and measure the frequency and time frames that the initial enthusiasm of new hires collapses, why and what can be done about it. 

Finally, we can use surveys to understand when changes in employee emotions occur that leave them more susceptible to the approaches of other organizations (or actively searching for new opportunities) and why. In my experience, more work can be done with data proactively such that these changes can be anticipated, and the root causes addressed prior to employee exit.

Join the Linkedin People Analytics Community group (~700 strong)

Connect with Mike West on Linkedin

People Analytics Question and Answer Series

What is People Analytics?

What is not People Analytics?

Why People Analytics?

What is the history of People Analytics?

What are the key questions of People Analytics?

What is the actual work of People Analytics?

Lean People Analytics Series:

Introducing Lean People Analytics

Getting Results Faster with Lean People Analytics

The Ten Types of Waste in People Analytics

Making a Business Case for People Analytics (with the three A's of Lean People Analytics)

Making a Business Case for People Analytics at Your Company (Get Started Guide)

Fueling People Analytics with Feelings

The Five Models of People Analytics

Talent Acquisition Analytics Series

Creating Competitive Advantage with Talent Acquisition Analytics

Measuring Employment Brand and Attraction

Measuring Candidate Experience

As always, I wish you and your tribe well! Happy insight harvesting to you. Mike West 

Michal Rovner. Border #8. The Met Art Museum. Open Access Collection.