In the creative workforce, good is very subjective. Fortunately for software development it's significantly easier to measure. A good developer completes a task defect-free (also known as, "good") and they do it quickly. Designers are a bit harder to quantify, but can still be measured based on user interactions and satisfaction. It's a large field to cover, so for the purposes here, I'm going to focus mostly on developers but the techniques and messages are applicable in any creative field.
Most people find their comfortable balance between fast and good, but there are exceptions. People can be exceptionally bad or good. Someone may be extremely slow and have many defects, and on the other side are the legends. A small percentage of people are fast and good. They have great reputations and are very selective about who they work for, with, and where.
The Traits of the Top Performers
Top performers are, obviously, highly desired and there is heavy demand to acquire them. This is more than just performance, hiring top performers means more gets done with fewer people. Between an average developer and a top performer there is a 2.5 times difference in performance! Two developers for the price of one! Jeff Atwood has a great post on this skill disparity, inspired by the book Code Complete. The post, and book, were written in 2004. It heavily references a study that ran from 1986 to 1988 called The Coding War Games. We have been able to describe a top performer, but we have consistently been at a loss to bring those traits out in others. In fact, current office and behavior trends show we don't know how to apply that knowledge.
First revealed in the 1980s, and backed by more recent findings, we learned that the environment in which creative people work contributes to their performance more than any other singular trait. Years of experience? Doesn't matter. College attended? Nope. Degree? Sorry! Two things matter the most: the individual's personality and their work environment.
A great developer is 10 times better than a bad developer, regardless of the years of experience or degree they possess. Even an above average developer performed over two times better than a median developer. If there is no strong correlation between degrees, colleges attended, years of experience where it the correlation? That link is the work environment. It is possible to increase the average performance by significant levels with small environmental changes.
Can we create top performers?
Create isn't the right word, but we can create the environment in which people can thrive. The Coding War Games, the study mentioned above from Tom DeMarco and Timothy Lister (along with a book they wrote) found that:
"Sixty-two percent of the best performers said their workspace was sufficiently private compared with only 19 percent of the worst performers."
Performance was clearly defined in this study, the work was timed, reviewed and defects were discovered. A third of the developers in the study turned in the work with zero defects, and they did it faster than the average. The top developers are fast and good. When surveyed, they tended to answer in the same way about their environments. The numbers suggests a strong environmental link. But a critical question must be asked, does the environment cause good work or does it attract good workers?
In the Coding War Games, each participating company has two developers except they don't interact. The developers remained completely isolated and work entirely separated. Even isolated, there was only an average 21% difference in performance between participants from the same company. This means that if you have one good developer, chances are you have another good developer. It's easy to credit HR and recruiting for being awesome, but the reality is that good developers are very selective with where they work. People want to be top performers.
The Environment Matters
In the top 25% of performing companies, employees had more space (78 square feet to 46 square feet) and 76% are able to divert their phone calls. The most critical statistic for me is interruptions. In the top performing group, only 36% said they were interrupted frequently. This is almost entirely inverted for the bottom percent. 76% of the worst performers said they were interrupted too often. Turn those phones on mute, close the doors.
How to get the best people?
DeMarco and Lister cautioned at the end of The Coding War Games summary that the work environment correlation is suggestive and the study doesn't prove anything. I agree, their study doesn't prove anything. Fortunately, more recent studies do; and even if we don't have conclusive proof but only a handful of correlative facts we can still make positive changes.
If there is a strong correlation between workplace privacy and control of one's time and thought, explore that further. Good employees flock together, and most likely flock towards organizations that have positive, healthy environments. Many studies have offered proof of how strong the environment can impact performance, productivity and happiness. I've read those, and simplified them (probably offensively so), and distilled them down into the top three steps to creating better environments.
Step 1. Stop interrupting!
Interruptions, whether a person popping in to say hi or a phone call, can absolutely destroy productivity. As I was writing this, the New Yorker just published an article on the dangers of an open workplace, citing studies from the last few years.
The largest performance drops are caused by the environment, and they affect everybody. This means a decline in attention spans and follow-through. Why do people lose their train of thought and get distracted? Interruptions! Gloria Mark, from University of California Irvine, found that workers get an average of 11 minutes between interruptions. Guess how long it takes to resume work after being interrupted? 25 minutes. Ouch. We'll be interrupted before getting back to full steam. No wonder we have short attention spa…
In 2013, the New York Times worked with researchers from Carnegie Mellon to test defect-rates when people are interrupted, or even expecting an interruption. When interruptions occur, defects increased by 20%. More defects means unhappy customers, turning into unhappy customer service, and then unhappy developers. Maybe we should augment Test Driven Development with Distraction-Free Development?
Interrupting causes more damage than we ever thought. It seems every study done shows even more of a negative impact than before. Fast, high quality work can only be reliably produced in an environment where, as a cultural expectation, interruptions are rare.
Step 2. Grant privacy
A 3M study (granted, they want to sell their screen protectors) showed that there was a significant drop of 50% in productivity when people were concerned about their visual privacy. Fortunately, since I don't trust paid studies, psychologists from University of Tennessee at Knoxville ran a similar study, which confirms that job satisfaction and performance increase with privacy.
If people are distracted due to the fear of being judged by what it looks like they're doing, versus what they produce then productivity and satisfaction will suffer. Yes, this means that sometimes (often times) developers will not be typing and staring at code; judge by output not by appearance and give people a private space. Let them. Embrace that notion that writing code, or any creative work, is not typing for 8 hours a day. Sometimes it's not typing anything in a day (Oh, how I hate those types of bugs!)
It's unreasonable to think that offices can be rebuilt easily to grant privacy, but think of smaller ways to achieve this (3M would love for you to consider their screen protectors, but I wouldn't consider that a viable option in most workspaces; I want the option of sharing!). The reward is worth it; boosting productivity 50% with no ongoing expenses!
Step 3. Ask & Listen
The thing that these studies and surveys did that most companies don't is simple. They ask people how they feel. I worked for a company that claimed to have a very liberal time off policy, yet every time I tried to take time off I was met with too many obstacles. This was not the company the CEO wanted to create but it was the culture. If I were asked if that was true then he would have learned. He never asked, I never told, and he never learned.
Find the key questions that matter, orient them to understand the perspective of the workers. One of the best habits from 7 Habits of Highly Effective People is, "Seek first to understand, then to be understood." It's fitting here. Also, don't argue about perception; perception is based on opinion.
If the majority of people do not feel they have uninterrupted blocks of time, privacy or even being listened to it is up to the leadership to learn this and then change the culture. The ideas don't have to be radical. Schedule Maker Days and tell people they can close their doors, or buy everybody headphones (the universal, "Do not disturb" sign). Also, when asking people's honest opinions make sure they report anonymously. We've already seen that privacy is very important. Dishonest answers will just go from bad to worse; people will continue paving the road to workplace hell with good intentions.
- Embrace an interruption-free culture, maybe starting by introducing the increasingly popular Maker Day.
- Ensure people have a privately owned space, but a comfortable and social open area they can voluntarily enter.
- Promote good and fast development with best practices starting from the environment and culture. Fix issues here before trying technology solutions such as Behavior or Test Driven Development.
- Stop the open floor plans. Stop the interruptions. Give privacy. Start being better.
To greener pastures,
Cover photo from Flickr user Michael Brown.