Stigmergy, collective intelligence and the open office
Last month I read the Bernstein & Turban paper on open-plan offices, where the researchers tracked two offices that adopted an open plan as part of a ‘war on walls’. They gave some workers a sociometric badge to track IRL interactions.
Their strongest conclusion is that face-to-face meeting time goes down and number of emails goes up. There’s also a cathartic, read-it-and-weep discussion section citing a bunch of research about overstimulation in the open office and tricks workers use (for example, headphones) to isolate themselves.
Their footnotes range widely across themes that I care about, and I’ve spent my last few weeks of commutes going through a stack of them. Highlights below.
This is a topic that more people are interested in post social media, but also means different things to different people. A survey on collective intelligence papers hits the crowd favorites like emergency and stigmergy but my favorite paper from this category is the collective intelligence factor in small groups.
Psychologists circa 1900 discovered an individual G factor that predicts mental performance across lots of categories. If you think about it this is remarkable and weird. This small groups paper is claiming that groups also have a simple intelligence score predicting their ability across different tasks.
If you believe them, the main components of the score are:
- ‘reading the mind in the eyes’ (a test developed by borat’s cousin, who is a serious neuroscientist)
- conversational turn-taking
- % female of the group (but the researchers hedge that this may be correlated with the others)
I don’t know about mind reading but turn-taking is something we can easily do better.
The % women finding (whether or not correlated by the other two factors) is interesting as well for its implications for hiring and labor law.
Bernstein, one of the open office study authors, did his dissertation on how transparency and privacy affect an organization’s ability to be productive.
Haven’t read this one yet, but I asked a bunch of manager friends about it. One is in an open office and told me his day is spent figuring out how to get people in a room to communicate with them. Another is at a place that has migrated most people to small offices and said that conversations are easier and his messaging volume has gone down.
I totally believe this. Open conversations often lead to pile-on and conversations get derailed like a government spending bill. Everyone wants to air their priorities.
Privacy in the personal sphere is morally ambiguous. Protecting and enforcing privacy in public institutions and the legal system has been harder than I expected. It’s refreshing to see a take on privacy that has wide support.
In theory the breaks in interaction paper belongs in this section – groups were better at solving ‘rugged’ problems when the group couldn’t share answers on odd steps. Their theory is that the top performers are brought down by the group.
Small group sizes also matter – the ‘wisdom of crowds’ effect is mediated by correlation of information across observers.
Knowledge worker productivity
I came across citations for ways to measure ‘innovation productivity’ but I didn’t read as deeply here and I didn’t understand what I read; my takeaway is that nobody knows how to measure knowledge worker productivity. Most of the papers in my set wanted to add a productivity analysis but didn’t have good metrics to use.
TFP, the widely-used economic measure of productivity, has never pretended to be what it sounds like – it’s purely a monetary measure.
But modeling productivity is critical to making sense of changes in the wealth distribution and whether technology is a red herring here. Piketty, one of the better-read economists of the new century, thinks technology concentrates wealth at the top (if I’m understanding him). ATMs haven’t put bankers out of work but we think the next phase of automation will?
Future of companies & work
Given all the above (and my working experience) I think modern work culture is adopting bad tools and practices.
There’s an Atul Gawande article on rigid forms leading to doctor burnout. The effects of burnout are hard to measure. You work more hours so your boss gets a better deal on you, but the work you do may have negative value for the team / company.
We’re structuring knowledge work in ways that leads to constant interruptions (which for my purposes is more than twice per hour). If you’re a social insect, stigmergy is how you use the environment as tool for external memory & computation; in our case it’s not serving our needs. Our communication tools don’t push back against overstimulation.
I see an opening for companies that will:
- Help teams measure abstracts like productivity, burnout, other abstract qualities
- Use these findings to attract good candidates; people are tired of having their time wasted
- Build third-party tools to do this and own the next generation of the ‘productivity platform’ (ms office, g suite, slack, …).