Overestimating and Magical Thinking

On a typical Saturday morning, it is easy for me to believe that I am at least three people. As the entire weekend yawns out before me, white and clean like a freshly stretched and gessoed canvas, in my mind’s eye I will paint on it three distinct images. There is one of me cleaning the house, organizing a closet, and finally filing that paperwork. There is another of me working outside in the garden, planting, tilling, weeding, hoeing. And there is a third of me having fun with my family and friends at a movie theater or dinner party, and connecting with my community. 

The trouble is, that constitutes at least three potential paintings, and only one canvas.  Unlike the painters of old, I cannot finish one scenario, scrape off the paint, gesso over it, and create another. To do so would require some serious magic or fantastic technology: time travel, the ability to go back to the beginning of the weekend and start over again with a whole new me. So invariably, I find myself at the end of the weekend, disappointed that I did so little of what I had imagined I would do.

I am not alone in my magical thinking about what I can achieve in a given time frame. Indeed, this nearly universal psychological phenomenon has a name: optimism bias. In general, optimism bias is a mindset that assumes that good things will happen to us, and bad things will not. In the realm of project management, it is re-dubbed planning fallacy: the belief that the time it will take to perform future tasks will be short, that interruptions and problems will not crop up, and that whatever complications occurred in the past when performing a similar task were exceptions, rather than the rule.

When I first started working with my client, K, a sole proprietor, her optimism bias was alive and kicking. She was new to the idea of estimating task size based on points, so we began with her smallest task, assigned it one point, and in true Scrum style used Fibonacci numbers to estimate the relative sizes of the rest of the tasks. In her first week-long sprint, she estimated that she could complete 50 points. Over time, we would learn that each point took her about an hour so, unbeknownst to us at the time, she was effectively estimating that she would be able to complete 50 hours of work in that first week.

In an ideal world, one where no problems occur, where no clients calls interrupt the work flow, where sleep is sound and one wakes up early and refreshed, where healthy meals are readily available and take no time to prepare, and mid-afternoon fatigue never sets in, a fifty-hour workweek may have been possible. Not sustainable, but possible. In K’s real world, where client calls are frequent, where mental blocks are inevitable, and where she is also a wife and mother with many chauffeuring duties whose son arrives home from school in mid-afternoon, a fifty-hour workweek is not even remotely feasible.

K’s optimism bias was compounded by the fact that she caters to many clients at the same time. When we started, she had twelve clients, and she estimated that she could work on the tasks of five projects during that first week. She started out the week like I start out my weekends: by imagining herself diligently working on the tasks of one project, then erasing that image and imagining the same for the next project, and so on. Five paintings overlain on a single canvas. Five simultaneous versions of K.

Magical thinking.

The consequence of K’s magical thinking was twofold. She was not only disappointed with herself at the end of the week, she was disappointing her clients. She would over-promise and under-deliver.

She does that a lot less now. Over the last six months that we have worked together on weekly sprint planning meetings and daily scrums, I have collected a lot of data on her velocity (number of points actually completed per sprint). One main insight gleaned from that data: with all of her interruptions and other life’s duties, she averages about 20 points per week. She still estimates that she will complete about 50% more than that at the beginning of each week, but she trusts me to reel in her expectations.

As a result, she tells me that I have improved her project completion timeline projections from 30% accurate to 90% accurate. Thus, she can communicate more realistic expectations to her clients.

I once heard that the only solution to planning fallacy is data, and this is certainly true in my experience. By keeping track of what K’s real-world velocity is, and continually reflecting that reality back to her in every sprint planning meeting, we are able to moderate her optimism bias.

Living in the real world is not as bad as it might seem. I recommend it. We may not end up with three paintings at the end of a sprint, but we — and our clients — will be a lot more satisfied with the one image we have painted. Both we and our clients will have faith that we can do what we say we will do. Their confidence in us will be well-grounded and, not insignificantly, so will our own confidence in ourselves.