Debugging and Optimizing the Modern Workplace

The name of my talk is “Debugging and Optimizing the Modern Workplace”. When I emailed with Sharon yesterday, the first thing she told me was “I know from my own clients that they bring in technology hoping to make things simpler, but often end up making things more complicated and stressful.”

I should begin with a little about myself. My name is Ezra Butler, and I am an interdisciplinary strategist. I consult with companies in the fields of solutions architecture, the intelligent workplace, and creativity. I studied for a bachelor’s degree in computer science and a master’s degree in history of comparative religion.

I left computer science, because while I really loved the algorithmic and programmatic thinking, I really hated the debugging. I don’t know if you know anything about coding, but if you have a misplaced squiggly bracket or a missing semicolon, the code doesn’t work, or even worse, it doesn’t perform correctly. If you ask any developer, something like 80–90% of their time is spent debugging code.

If debugging is figuring out all the problems with the code, optimizing is making sure that it works in the most efficient way. If you develop the greatest mobile app in the world, but it’s too big for an iPhone, then you have a problem. Something can be bug-free, but not optimized.

Both debugging and optimizing start with asking the right questions.

When most people use words like “debugging and optimizing”, they are usually talking about using technology and code, but it’s not necessary. Today, we’re not.

Today we are borrowing the terms “debugging and optimizing”, because the methodology is the same, no matter if we are using whiteboards, notecards and paper questionnaires or a solution that does it for you with a Google Form and the Salesforce app. Today we are learning to ask the right questions.

Technology is not a magic silver bullet, and just because something is “smart” doesn’t make it good. Just look at the lawsuit recently filed against Amazon for using computers to track warehouse workers’ productivity and algorithmically terminate them without human intervention if they fell behind pace. Do any of us want computers to decide the fate of our employment?

On the flip side, if you look at all the human biases evident in the hiring processes, with the subconscious sexism, racism and other forms of soft bigotry, there may be a space to bring an objective, non-biased voice to the hiring process and human resources as a whole.

HR is broken in many ways. When you think about it, our employees are our most valuable and expensive investments. The cost of hiring, management, perks and benefits, healthcare, 401k, and everything else adds up.

We require extensive analytics, tracking and A/B testing for marketing campaigns for a fraction of the salary of any employee. Yet, we allow people to book hour-long meetings for 20 people, considered by all the attendees to be a huge waste of time, which effectively cost the company thousands of dollars, and no one blinks an eye.

I was speaking to an employee of a massive media conglomerate the other day, and he explained to me that he receives more than 100 internal emails a day, most of which are completely irrelevant to him, but he still needs to spend 20 seconds perusing each one to ascertain if it holds any relevance to him.

These may be edge cases, but I would contend that we can all find some similar time waste in our own workplaces.

More than 30 years ago, researchers estimated the cost of workplace stress to be more than $300bn a year. In the interim, work became a 24/7 job which is always-connected and always-on, for so many of us. The 9–5 workday of the 1950’s has fallen by the wayside. Business has become global, with employees operating in every time zone, introducing new challenges like middle-of-the-night conference calls, increased travel, and last-minute deadlines.

In 2017, France actually passed a “right-to-disconnect” law, requiring companies to allow employees to ignore their smartphones and work email when they leave the office at night. A French politician said that the employees “remain attached by a kind of electronic leash. The texts, the messages, the emails — they colonize the life of the individual to the point where he or she eventually breaks down.” C’est la vérité.

New estimates place the cost of employee disengagement, stress, and distractions well over $1T a year. “Our work can literally make us sick. Long hours, impossible demands from bosses and uncertain job security can take their toll on our mental and physical well-being, leading to stress-induced aches and pains and anxiety. In extreme cases, the consequences can be worse — heart disease, high blood pressure, alcoholism, mental illness.” wrote Michael Blanding for Forbes in 2015.

Silicon Valley companies figured a different solution. They would offer free meals, laundry and childcare services, happy hours, and the infamous ubiquitous ping pong tables, anything to keep employees in the office late at night.

I was present at a developer meeting for a startup a few years back. One of the developers, a young woman in her mid-20’s, raised her hand and said that she was exhausted from being forced to work late, until 9 or 10 at night, whenever they had a new feature to launch, and then being required to come in at 9 am the following morning for the mandatory standup meeting. The head of developer told her “You know that you have unlimited vacation days. Just inform your manager that you need to take one.” She appeared deflated.

I went over to the CEO immediately after witnessing this interaction and made a simple suggestion. “Perhaps if someone works until after 9 o’clock at night, they don’t have to come in until 11am, without having to ask someone for permission.” They obviously could come in first thing in the morning, but there is the accepted practice that it’s not required.

In the title of this talk on the website, it included the phrase “Using psychology and technology to reduce stress and make your office a much happier place.” Let’s talk about psychology first, because technology seems to be the reason for so many of our problems.

One note on optimization: We have to optimize for something. In programming, it could be for file size, quality or cost. We’re not going to optimize for productivity, efficiency or billable hours. We are going to optimize for the 3 A’s.

The first A is for examining the level of employee Agency. Agency is the ability to act, whereas autonomy is ability to self-govern. Psychological research shows that the more agency that someone has, the happier they are. That developer we previously spoke about had the illusion of agency, but she didn’t really have it. When you have to ask permission to do something, or even be required to give a “head’s up to management”, you don’t have agency. Neither do the employees in the Amazon warehouses beholden to their robot overlords.

When employees are given the agency to take risks, execute on ideas, and try new things, they tend to come up with more creative projects. Google’s 20% project comes to mind. They allowed employees to take up side projects on company time. Sure, it’s not all altruistic. The idea that the employee would be working on at home after she left work is now considered intellectual property of Google.

Debugging the process is first recognizing that this is a problem and then identifying potential solutions to solve it. But the truth is that we don’t actually know if it will work. I don’t know if you implementing a 20% project with your company will work, or what the equivalent in your line of work is. Debugging is like the scientific method, where we first have to come up with a hypothesis and then test it, and see if things worked better. And then we perfect it with a feedback loop. And that is the optimization part of the process. How do we make it even better.

But when we try — employees notice. They see that we are trying to do something to make their work life better. This is a version of the observer effect in physics: you can’t measure something without having an effect on it.

The next A is for attention. To steal a term from technology, we all have limited mental bandwidth. Research shows that there is a true cost to task-switching. In other words, if I’m working on a report, and you come up to my desk and ask me a question, it will take me a while to mentally return to the work at hand. If you send me an email or a slack, it will have a similar effect. The minor interruption has major implications, to the tune of an estimated hundreds of billions of dollars wasted a year globally.

As an aside, I’m not a fan of statistics. They are effectively meaningless. Think about yourself. Think about how frustrated you get when you are trying to do some work, and you keep getting interrupted. Or when you are finally getting into a flow on one project, when you are pulled into a meeting or onto a call regarding something completely different.

Let’s go back to our previous examples of the 20 person hour-long meeting, which forced all the attendees to stop their work 20 minutes before the meeting started, and then to decompress for a half-hour after the meeting to get back in the work mindset. And this is to say nothing about the meeting prep or the followup. Or the media employee who is getting interrupted with email every 4.8 minutes, on average, and needs to take a half minute to get back to the work at hand. How could anyone get any work done under these conditions?

And the third A we want to look at when debugging a company is the ability to express emotion, and more specifically, what the accepted reaction to a negative experience is. This could be where someone loses a big sale or where a partner is sexually inappropriate with an associate. I just read an article the other day about how minorities consistently are confused for each other in the workplace, and are often criticized for making a big deal out of being misnamed. And this does not even touch on how a trans person feels when they are misgendered or deadnamed in the office.

At the very least this limits their professional development.

Does the employee feel comfortable expressing emotion? Does the employee have an outlet to be heard? Is there a system that actually deals with this, or is it just theoretical?

In Expressing Emotion, Drs Eileen Kennedy-Moore and Jeanne Watson write about an interesting paradox: Expression of negative feeling is both a sign of distress and a possible means to coping with that distress. Without the ability or tacit permission to express oneself, one has to find other outlets. Drugs, alcohol and repression become outlets to deal with the stress, leading to hungover, irritable, and disengaged employees. Going to Happy Hour every night becomes an ironically named coping mechanism.

After meeting with a leading epidemiologist who helped track the spread of Ebola and other highly communicable diseases, he commented “wow, stress is a contagion in the workplace.” All it takes is one person in the office to come in a foul mood, and it spreads. See how much work gets done the morning after the local sports team loses the championship.

Agency, attention, and the ability to express emotion. If an employee doesn’t have these three A’s, your company has underlying issues. It doesn’t matter how many ping-pong tables and classic arcade games you have.

Now we know what we are looking for. How do we figure out if we have these problems? And if we try to solve them, how can we gauge if we are getting better?

Especially with regard to the third A, the ability to express emotion, people are not very forthcoming with honesty. Because they see what happens when someone speaks out. It seems like it is easier to get rid of a squeaky cog, than to oil it.

Current engagement surveys are lengthy, expensive, infrequent, and riddled with biases of all sorts. The simplest to explain is something behavioral psychologists call satisficing, which basically means giving the sorts of answers they know the questioner wants to hear. If you are asked to fill out a survey filled with rankings from 1 to 10, how often will your answer end up on one of the ends? What is actually the difference between a 6 and a 7? Is your 7 different from my 7? And if it doesn’t matter which it is, then why should I waste my time answering it?

Researchers from Northwestern University and the University of Toronto recently gave out two different surveys to undergraduates about what they thought about their lecturers. In one set, it was on a scale of 1 to 10. In the second set, it was on a scale of 1 to 6. Under the 10 point scale, 31.7 percent of the male ratings were a perfect 10, the most common result for them, while only 19.5 percent of the female lecturers received a 10. When ranked on a 6-point scale, 41.2 percent of men scored 6, while 41.7 percent of women scored the same.

Now what happens if you give a survey right before pay day or immediately after pay day? In their book Scarcity, Drs Sendhil Mullainathan and Eldar Shafir show how having less money has a negative effect on cognition. If you are a millennial employee living from hand to mouth and can’t afford lunch that day, how do you think you are going to think about the company?

What about when you survey employees on a team they just joined, or one which they have been part of for several projects? What about after they receive their bonus or return from holiday? How would any of those change the responses?

Spot checking for the 3 A’s doesn’t work. You can’t simply assume that because one day everything is alright, that it’s always that way. You have to continuously track for them.

Many companies are now using simpler automated weekly surveys, which solve certain problems, but others remain. The most basic is: if you do not feel comfortable expressing emotion in the workplace, would you honestly respond to a questionnaire which could be easily tracked back to you?

For a little more than three years, I worked on a project I started called Quantified³. The goal was initially to reduce stress in the workplace, but it led me down a rabbit hole figuring out how. I had begun with research I had done nearly a decade earlier, when I had a startup called Happy | Not Happy who was a finalist in Seedcamp Tel Aviv. We had built an app for psychologists to prescribe to their patients to combat the recall bias. The recall bias is that we most readily remember the most recent thing that happened to us. Whenever a patient felt happy or anxious, they would just press the correct button and it would store the date, time and location. When they would next visit their psychologist, they would use it as a jumping off point for conversation. So instead of saying “I got into a fight with my boyfriend last night and I’m still in a bad mood.” They’d be able to remember to cover more extensive, recurring issues as opposed to fleeting ones that were top of mind.

I wondered if a similar thing could work in the workplace, but in an updated way. As I built, tested, and rejected new models, I delved deeper into how to lower the cognitive load, and to increase the expression of emotion. People began using phrases to me like “that meeting really had me pushing the not happy button.” We focused around meetings and interactions, giving people the ability to anonymously and securely debrief from every meeting. What went well? Did the meeting achieve its agenda?

Besides for the positive benefit of being able to express emotion in the moment, mentally deal with it, and then put it aside, we began to design functionality that would benefit the employee personally. For example: each one of us has a different social bandwidth. Some people could sit in 10 meetings a day and walk out energized. I used to have a business partner that I needed to schedule all meetings with him right before I would have a three hour window free to mentally decompress. The goal of the algorithm was to figure out how many meetings a day were optimal for the mental health of each employee and during which part of the day. The next stage was to suggest times for meetings without explaining the rationale. So if Sharon’s a morning person, and Wednesday is my optimal day of the week to talk about this project, the algorithm would suggest Wednesday morning at 10:30am to have this discussion.

As I said before, technology is not a silver bullet. At the very least, we need the three A’s: Agency, attention, and the ability to express emotion. We should be using technology to ensure that those three things exist. A simple calendar notification could say that Roberto is out of the office today, without being required to inform a superior. A do not disturb autoresponder on emails sent to other employees could inform them that Sally is working on a project and will be unavailable for a few hours, even though she is alone at her desk. And some sort of anonymous message box could provide a way for people to express emotion. Or maybe there are three better ways. But you won’t know if something works until you test it. And in order to measure that, technology could be of assistance.

When I first looked up Boye & Co yesterday, I learned that all the members are “people who are working to solve the same challenges, but who might have found altogether different answers.” It is my suggestion that we begin with making sure that we are asking the right questions.

Thank you very much for listening.