I’m being onboarded as the new reviewer for the Workflow manager epic. As part of this, I went ahead and read the book, too.
I think Newport’s suggestions of meetings make a lot of sense, though they need to be contextualized (he made it pretty clear that meetings can be a big timewaster, and in fact are sometimes used to paper over work and give a sense of accomplishment.) Unfortunately even when well employed, they’re less useful in our case than for others.
A particular tricky lesson to pull from the book is that asynchronous communication methods, by default, tend to be inexpensive to send and expensive to respond to. They also encourage ‘always-on’ behavior where you’re paying attention to pings that will break your concentration, and feel distracted if you haven’t answered them. This distraction is deeply primal and has origins in our evolutionary history-- so it’s not something so readily ignored.
This causes a ‘tragedy of the attention commons’ where it’s very easy to break someone else’s concentration. Email acts as a ‘catch-all’ for coordination activities that have no formal intake and output processes. The net result is that everyone is unable to focus on real work because everyone is focusing pings.
This presents a unique problem for OpenCraft, as a company that is all remote. In fact it’s a problem that I imagine has gotten worse with time, as I’ve been around long enough to see how some things have changed.
We started by using Trello. Trello is highly celebrated in the book, and for good reason. The card-based method of organizing conversations, and being able to engage when you’re ready, helped a great deal in cutting down the amount of asynchronous chatter. As I recall, we switched to Jira due to a few technological limitations, especially surrounding time tracking and billing.
Jira, however, sends far, far more emails. It’s also a tool that adds a lot of friction by trying to be everything and do too much at once.
Looking at the pain points in our process, a lot of them involve the back-and-forth through Jira as we plan sprints, and while I’m usually not involved directly in help@ and urgent@ email threads, I do find myself actively engaged in contact@ emails, which often aren’t tied to any specific ticket.
Reading this book leaves me full of ideas, some of which we’re not in a position to execute on any time soon, but which are fun to think about for long-term possibilities:
- Could we migrate our contact lists to a ticket system? One that can tie threads directly to tasks for us, and put the history right in our app, in a card, rather than piling up in our inboxes?
- We do have a policy of 24-hour turnaround for notifications from clients. How do we ensure that all items are responded to within their time limit without having a big inbox pile?
- Where are the key attachment points for plugins in the workflow software? This is important to consider-- because it will affect what patterns of automation we could potentially support.
- We’re trying to avoid building a giant monolith that does everything like Monday or Jira. But we do want to make several smaller apps that can work together. Doing so will likely remove a lot of context switching by providing oppportunities for automation. Are there particular workflows and interactions between the apps we want to consider? This is similar to point 3, but more concrete.
I’m also excited to work on this project. Not only is it a really cool problem to work on, it’s one that I’m certain I’ve already written a lot of software that may be useful for. To elaborate, Artconomy has a great deal of minutia-management that these apps are going to need. Rather than writing it all from scratch, it would likely be cheaper for me to port this code, add documentation, and turn it into libraries our apps can use. So I look forward to that possibility. Some of the things include the websocket code, the form and state handling, notifications system, and potentially even the financial calculations. I already know that the unique shortcode handling function will be useful, and that one’s a library that’s already ‘ready to go’ here.