Office Climate Change
There are a million ways to lose a workday, but not a single one to get one back.
As I have written about previously, a common belief in the software community is that overtime is part of being a software developer. However, the prosperity that the software industry has is not due to the quantity of software being produced, it is due to the quality of what is being produced. So perhaps some of these extra hours are not being spent on quantity, but on quality.
How often have you heard someone say they do their best work before 9am, or in a late evening, in the times where the office is more empty? People even work from home in order to focus on the task at hand. This is a very strong sign that the office environment is not serving its full purpose. Often, problems with the office environment are treated as out of everyones control. True, usually a select group of people have control over things like office layout, but it is not beyond the team and manager to convince these people of the benefits of changing the office environment for the better.
A very useful study was done by Tom DeMarco and Timothy Lister which they deemed the “Coding War Games”. This put 600 programmers against each other attempting to complete programming tasks to a high quality as quickly as possible. Not only was it a lot of fun, a large amount of data was collected, and each member could see how they ranked against the pack. However, the reason I mention this is that they required that each person perform the task independently from another person in the same organisation, so at least two people were always entered per organisation. The results of all the individuals followed the expected outcome of any performance metric, the best outperformed the worst by about 10:1 and the best outperformed the median by about 2.5 times.
What the results showed that you might not expect is that language, years of experience and salary had little to no impact on the performance of individuals. What did matter was who the other person in the organisation was. The pairs, who had to work independently, not helping each other, performed an average of only 21 percent different. This is important because for the overwhelming majority, these people performed their work within the same environment, suggesting that people within the same organisation tend to perform similarly, that the best people work with the best people and the worst with the worst.
It should also be noted that Harlan Mills did a study which showed that there is a 10 to 1 difference in productivity between organisations. This should make you realise that the office environment and the culture which determines who joins and stays at companies, is the dominating factor when it comes to productivity.
Most companies provide developers with crowded, noisy and interruptive environments. In a lot of cases, this is enough of a reason for good people to move on. In the “Coding Wars”, each programmer filed out of a survey on their office environment and the results are striking. In terms of space to work, top performers have 78 sq ft, whereas low performers have 46 sq ft. Top performers said their workplace was acceptably quiet 57% of the time, as opposed to 29% for the low. 62% of the high performers said that the environment is acceptable quiet, with low performers reporting only 19%. The question “Can you silence your phone?” was answered positively 52% of the time and an abysmal 10% by low performers. One of the biggest differences was the answer to “Can you divert your calls”, with a whopping 76% of top performers saying yes, but only 19% of low. Finally, interruptions, it turns out the top performers are interrupted often 38% of the time, but low performers are interrupted an incredible 76%.
It’s a lot of data to take in, but it is extremely clear that there is a very close correlation between work environment and productivity. This does not necessarily mean that improving the workplace will improve the current programmers productivity, however, it does mean that in the long run, improving the workplace will mean that the best people are hired and retained more frequently.
Granted, the low performers probably aren’t employed by people who think that the office environment is optimal. More likely they think it’s the cheapest option. However, they aren’t doing the cost/benefit analysis on the spent money, they are simply looking at cost bottom line. At the end of the day, a favourable environment for a developer is a very small percentage of their overall salary. Ranging from 6 to around 16 percent. So when we consider that a programmer might cost 20 times more than his workplace, it should be seen as quite a high risk to cut costs on their environment, since the return on investment is so high, no?
Without warning, and relatively overnight, the open-plan office has hit us and been taken up in almost every technology office on the planet. There was very little, if any, evidence that a open-plan offices improved productivity. There was also nothing to say that productivity wasn’t impaired by open-plan offices. Articles on the benefits of open-plan were produced everywhere, without any scientific backing whatsoever. Companies started taking up the new office layouts, and these companies became justifications for others to follow. How many times have you heard, well that’s how they do it at Google/Facebook/Instagram/Apple? As if we have any sort of concrete research on the productivity of individuals within such places.
So who should come to the rescue, and study the habits of the people who would be working in their newest office? IBM of course. They watched programmers, engineers, QA workers and managers, in existing workplaces and mock-up workplaces. All the way back in 1987, IBM decided that maximising productivity meant: 100 square feet of space per worker, 30 square feet of surface per person and noise protection from enclosed offices or 6-foot high partitions. The resulting office had about half of all personnel enclosed in one or two-person offices.
Why did they do this? Why did IBM go to the trouble of constructing such an environment? Because they concluded that people needed the space and quiet in order to perform optimally. A reduction in cost would mean a reduction in effectiveness which would result in a net loss in savings. There are other studies which confirmed IBMs findings, but IBM were the forerunner of this research.
In the “Coding War Games” only 16% of workers had 100 square feet or more, only 11% were enclosed in partitions or offices and there were more people in 20–30 square foot spaces than in the 100 foot spaces group. 61% of people reported that their workspace was not sufficiently private and 54% reported that they had a better workplace at home than was provided by the company.
Digging further into the data, workers who reported there workplace as quiet were one-third more likely to produce zero-defect work. So what happens in environments that are too noisy? People often book conference rooms or visit coffee shops to complete work. That’s right, they aren’t sat in their tiny allocated space but they still need to feel the accomplishment of work completed. People will do whatever it takes the make it happen, even if their office environment is against them. These sorts of things happen when a company tries to save money on the environment and doesn’t realise it is bleeding away multiples of the savings.
IBMs decision is a useful data point, the problem is, how do we actually measure the productivity of people in intellectual lines of work? They aren’t factory workers who can be reduced to number of stuffed animals constructed per hour, their work is fundamentally creative. Most organisations don’t measure their own output in any meaningful way, nor do they look closely at the costs very closely either. Turns out, we don’t have to come up with the perfect solution to this problem, we have absolutely nothing, therefore something is an improvement. There are plenty of schemes defined for software productivity measurement, there are even companies who will come in to assess the productivity of a workplace in comparison with the rest of the industry. Given that there is a 10:1 difference between the best and work performing organisations, it is simply unacceptable to not take any steps in order to work out how far up the scale you are and how you can improve it. The savings in productivity will vastly outweigh the cost of measurement.
Unfortunately, measurement usually leads to fear and and reduced job satisfaction. People do not like to be intrusively monitored, they don’t want their self-esteem at risk just for some study. Therefore, the management who are interested in the productivity cannot be given data on individuals. It has to be clear and both parties must trust one another, the measurement has to be done as a whole group or organisation, not on individuals, which quickly turns into a witch hunt. We always have to consider the people aspect of any problem, this is a key time to do so, no matter how control hungry the management is, they need to let go in order to reap the benefits.
That isn’t to say the individuals themselves can’t have their results given to them, perhaps in more vague terms than has been collected, such as areas of strength and weakness, with suggestions on how they can improve. As long as this is in strict confidence, people will regulate themselves, doing better on their weaknesses and driving home their strengths. After all, their self-esteem demands it.
Anyone in intellectual work knows about flow. The feeling of looking down to work and then looking up to realise that 3 hours has passed without any perceived effort being applied. It has been shown time and time again that it takes around 15 minutes to descend into concentration. During flow, you are overly sensitive to noise and interruption. Every interruption is another 15 minutes lost in getting back into the flow. This is a frustrating process, getting interrupted during flow is a sharp gear change which is not exactly pleasant and having to get back into the flow requires a lot of activation energy.
One of the problems is that managers are typically unsympathetic towards these problems, after all, they are more likely to spend their days in meetings, taking phone calls and catching up with people. They need to realise that the people who work for them need flow. Stopping flow costs money.
When a company tracks time spent on a task, they almost always track time at seat. Regardless of how many of those hours were productive. On top of that, almost always, a worker is expected to track the total number of hours they work each week. People will always add time where they can in order to meet the expected quota. This makes the task meaningless.
A more useful metric is measuring number of uninterrupted hours. People need to be aware that it is not their fault if they are being interrupted, the task is to measure the environment, not individual performance. Not only does this shift the data towards being useful, it enables a culture in which people think about the importance of flow and attempt to maximise it, protecting their flow hours. The organisation is suddenly interruption conscious.
So now we have some data, what can we do with it? We can measure the effectiveness of the environment, Simply work out uninterrupted hours / desk hours and we have the environmental factor. High performing organisations achieve up to 0.4. It is also important to look at different parts of the organisation and understand why the environmental factor is different, in order to improve on those areas. This is a very powerful metric, because a value of 0.2 means half as much productive time as 0.4, so each person is roughly twice as productive if the environment can be improved. It’s hard to argue that the cost is not worth it at that point.
What if a lot of people have 0 uninterrupted hours? Well, once people are faced with a big zero, they will become aware of the fact that it’s probably a problem. The culture can then focus on becoming more interrupt-free. It is a justification to hide away or shut off your phone for a couple of hours a day in order to be most productive.
This is going to sound controversial these days; but I think we need more doors. Doors allow workers to control noise and interruption. Think about how powerful that sounds, workers can actually decided when and for how long they want to be productive. Workers will maximise for productivity, if the organisation only functions well if they have their door open 5 hours a day, so be it, but allowing them to close it for other 3 will make people feel empowered, trusted, respected and produce immense amounts of value.
Is it too late? Well, it’s never too late. People just need to be brave enough to speak up. I think the vast majority of people would rather work in one or two-people offices, with the ability to close their door, not only because they would be happier (which in turn increases productivity more than anything else), but they will be able to focus more on the task at hand. Speaking up means that it is harder for management to keep cutting costs on the environment, if people are vocal in that they want improvement, it would look very poor if they started decreasing the quality of the environment.
It won’t change the world overnight, but more voices are better than less noises. Management will push back on the door, saying there are cheaper solutions and that people will not interact if they can simply close their door. One way that people address the problem is to listen to music. However, it has been shown, all the way back in 1960, that someone who claims they enjoy working with music will produce similar quantity as a person in a quiet space without music, but at the same time, that the quiet person will produce more creative work. Music will not effect the significant portion of work that is simply information processing, but it will effect creativity. Intellectual work is creative, it requires creativity in order to solve the tough problems, in order to see things in a different light, in order to have those Eureka moments. Simply listening to music does not address the problem at hand.
Another important thing to note is that we are not suggesting every person has their own office, two, three and four person offices make a lot of sense, particularly when these people are aligned as a team. This allows the interaction whilst also enabling team flow, an extremely potent thing.
Having said all this, what should we take away? Two questions need to be addressed; What kind of space supports your workers in order to make them comfortable, happy and productive? and What form of work space would make these workers feel best about themselves and their work. Think about the optimal working environment, it might be upsetting if you are a in a very poor one, but it is a useful exercise. We aren’t trying to get the company to knock down the current environment an build the utopia you imagine, but we are thinking about the stars in order to first reach the moon. In my next post, I will write about the key things to think about when creating an new office environment. For existing spaces, simply make your voice heard, try to get small change happening and you never know, the culture might shift in your favour and maybe one day you won’t have to listen to that guys loud daily afternoon phone meeting.