Dear future colleague, safety is our superpower

Timo Rantalaiho
November 11, 2021

Read time 4 min

I felt my heart beat quicker as I raised the cup to my lips to drag in another nervous sip of early morning production deployment coffee. The lines on my terminal rolled by. The script stopped the old version of our application, changed the symlink to point to the new one, and then started it up. This early, most of our users would be offline. But soon they would start flowing in again, and I was anxious to see if “This time it would work”.

After spotting the logline indicating the service was back up, I immediately alt-tabbed back to my browser and refreshed the front page.

No! It was still broken! How on Earth?

This was no time to check out the extra debug logging I had added to investigate the problem, so I quickly started the rollback script to bring the old, functioning version back up.

* * *

Dear future colleague,

I had always thought that in our kind of challenging knowledge work, it is essential to be able to both ask for and to give help easily. In programming and in delivering software systems – the fields most familiar to me – there are simply too many things for any single person to master thoroughly. Having friendly colleagues give and use help from each other gives us a boost that makes us deliver more as a group than would be the sum of the output of each of us alone. Their help makes 1 + 1 = 3.

Recently I was introduced to Daniel Coyle’s book The Culture Code: The Secrets of Highly Successful Groups (2018). In it, Coyle uses captivating stories to give compelling evidence on what makes some groups especially successful in their fields. The examples range from business to entertainment, from sports to warfare. 

A key concept that Coyle digs up from each case is shared vulnerability. When members of the group feel that they are free to expose their vulnerability and to react to the vulnerability of others in authentic ways, the group manages to function together exceptionally well.

This was a missing link in my thinking about the importance of helping each other. When I ask for help, I admit my shortcomings, which makes me vulnerable. 

As I was working in a good-humored group of friends, I had never paid it any attention – it didn’t even cross my mind that there was anything special about helping each other. Similarly, it’s easy to not think about tap water when it’s drinkable and tastes good.

In a blog post, Coyle further elegantly connects sharing vulnerability with a feeling of safety in “a behavioral model of culture.” He writes:

The three signals work together: safety and shared vulnerability reinforce each other, combining to create ever deeper connection; story provides the direction, the purpose. More importantly, this model provides the how — a way to solve problems.  

This concept of psychological safety was first introduced by William Kahn in his 1990 paper “Psychological Conditions of Personal Engagement and Disengagement at Work.” In a couple of interesting case studies, Kahn finds that the felt psychological safety significantly helps to engage with work.

Autonomous engagement is crucially important in the self-organized knowledge work we do. 

And Coyle shows that the safer we feel in a group, the more we can share vulnerability, and as a result, the more we’re able to achieve great things together.

Because of historical reasons, Reaktor has long been a safe place to show vulnerability – for pale-skinned, omnivorous, heterosexual cis-males born in the era of Kekkonen. 

For us to retain this strength as we grow, we must consciously work to make Reaktor safe for the increasingly more diverse group of our colleagues. Together.

* * *

The debug logging showed that the bug I’d been hunting for was caused by a peculiar difference between our test and production environment. In production, the database, for some reason, added blanks after a constant string of characters that it returned for a query. I had made a union and differentiated between the two different sources by constant values, but when I filtered the data coming from different sides of the union, I found no matches. This is because I was comparing "foo" with "foo   ".

At that time, we did not have automatic provisioning of servers. So both production and test servers were handcrafted, which was the most likely reason this problematic behavior didn’t occur in a test environment.

“So, good morning! How did the deployment go? Does the new front page work now?” Our customer had arrived at the office.

“Er, ehm, no, it’s… I mean, no, it still didn’t work. I think I finally figured out what’s wrong, though. But for now, I just rolled back the old version, with the rollback script.”

The red rising up my cheeks was cooled down by the warmth of their smile. 

“OK! It’s good that we made that rollback script, then!”

Dear future colleague: A series of letters written by Reaktorians. Come join us, as you are.

Here are some sources and further reading:

  • Coyle, Daniel 2018: The Culture Code: The Secrets of Highly Successful Groups
  • Coyle, Daniel 2018: “How to Think About Culture”, at Retrieved on 6th of October, 2021.
  • Kahn, William A. 1990: “Psychological Conditions of Personal Engagement and Disengagement at Work”. Academy of Management Journal. 33 (4): 692–724.

Never miss a post