Yet another of those ideas in Thinking Fast and Slow by Daniel Kahneman that leapt out at me as I read was
One of the significant discoveries of cognitive psychologists in recent decades is that switching from one task to another is effortful, especially under time pressure.
This seemed relevant to project work. Task switching is certainly costly in projects.
Before getting to switching individual tasks, it seems relevant to visit the topic of switching tasks mid-sprint. Switching tasks any time mid-sprint in agile (and indeed any tinkering with scope mid-sprint) has, in my experience, not worked well and this seems related to what Kahneman is saying. Others seem to agree about the costs of switching tasks mid-sprint. I enjoyed the exploration of this theme in the “Product Owner changes” section of this article and the “Changing Requirements” section of this one. Substitution mid-sprint is a slightly different proposition to change. I am all for substituting scope during the project. For example, where the team is a 3rd party supplier on a fixed price and the product owner identifies new scope mid-project, we have estimated the new scope and substituted out like for like stories. This is a technique that has worked well for me and I would recommend it. So substitution mid-sprint/project can work well but changes in the middle of a sprint carry a cost to performance and morale that has often been significant for me and my team. There are some reports to the contrary, and I’d be glad to hear of other’s experiences with this.
So, Kahneman is advising of the costs of asking someone to switch from one task to another before the first is finished. This is something that should be relevant to agile and waterfall teams alike. The articles I mention above discuss the cost of switching tasks mid-sprint, which, as I say, seems to have related overheads. Much of what these articles say translates to the cost of changing work packages mid-stage in waterfall land. Work packages often come with tolerances which, if exceeded cause project issues to be triggered but we are still left with the psychological cost of task switching. So what is this cost?
As will be familiar to you, it takes a certain amount of mental investment to understand a task and think through how the requirement is going to be met. Once the task is done, the back story and details can be forgotten but, until then, it all needs to be kept in mind. Blumar Zeigarnik demonstrated how we remember unfinished tasks more easily than completed ones. The eponymous Zeigarnik Effect that she identified goes some way to explaining the cognitive effort involved in incomplete tasks versus complete ones. The theory is that incomplete tasks take up mental capacity which is “released” once the task is complete. It makes me think of caching and flushing the cache in computing but this may be on the wrong track. Such an allocation issue with mental resources would go some way to explaining the drop in productivity from switching tasks. Perhaps an explanation is emerging but the Zeigarnik Effect is not the whole story.
I imagine the “significant discoveries” Kahneman is referring to is the body of research suggesting multi-tasking is not a great idea. While the American Psychological Association (APA) says that
Although switch costs may be relatively small, sometimes just a few tenths of a second per switch, they can add up to large amounts when people switch repeatedly back and forth between tasks.
My own experience (and that of my team members) is that task shifting carries a significant cost per task (not “relatively small”). The APA article mentions two cognitive costs of task switching, “goal shifting” (getting your head around the new task) and “rule activation” (if this is what I am going to be doing then this is what I will need to bear in mind). I wonder if this is what we project managers call the learning curve or something different. Since both of these need to be repeated in some measure when you go back to the old task, they are pure overhead.
Another component which I mention above is the blow to morale. Completing tasks carries intrinsic rewards. We experience these as this is experienced as positive physical sensations, as the brain releases neurotransmitters upon task completion (notably dopamine). Similarly, extrinsic rewards often occur in projects in the form of peer recognition and stakeholders showing approval and appreciation for completed tasks. Task switching means that the team member is now further away from the satisfaction of completing the task. While they were working towards that satisfaction on the original task, they will now not get that reward (intrinsic or extrinsic) any-time soon. Instead, they are now back at the start of something, so while they might have been hours from the satisfaction of wrapping something up, they are now days from that pay-off.
To summarise thus far, we have cognitive overheads like the Zeigarnik effect, goal shifting and rule activation. These are perhaps taking up memory and cognitive resource for the act of task switching. We also have social effects such as missing out on peer approval and stakeholder appreciation and the biological perspective of missing out on the positive sensation of neurotransmitter release in the brain, which is part of the mechanics of how we motivate ourselves. There is also the cognitive perspective on intrinsic motivation. All of these paint a picture of some of the main costs in time and morale that come from task switching. Incidentally, in writing this post I came across a surprising theory on the interplay between intrinsic and extrinsic motivation called cognitive evaluation theory. It seems positive feedback is not helpful where people are already self-motivated. Kahneman mentions that the cost of task switching is especially high “under time pressure”, which projects are almost by definition (and many projects extremely so).
Of course, project teams don’t have the final say on if the task gets switched and it may be worth it to the organisation despite these overheads. What seems important is for organisations to understand the full cost of task switching. It’s more than I thought it was.