Estimation

Inaccurate Estimations

When my wife and I had our first child I would consistently (and drastically) underestimate the amount of time it would take us to drive anywhere further than 50 miles or so. Often what I would estimate as a three hour drive to go 160 miles would turn out to be a four (or more) hour drive.

As an individual driving alone, I'm very much a "start to finish" driver. I don't speed but especially on highway trips I set the cruise control when I get on the highway at the start and turn it off at the last exit of the journey onto local roads, and do as far as I can at a stretch with some music on until biological imperatives kick in.

For a 160 mile drive to estimate my arrival time I would figure in at worst one rest stop and a few minutes of slow traffic along the way. I drove like this for 13 years before I met my wife and 18 years before our first kid, so my gut estimate for highway trips was essentially "a mile a minute, plus a few minutes" and it was reinforced in my brain as accurate for hundreds of thousands of miles of highway driving.

What I didn't realize as a new parent was that small children add a substantial amount of stops and those stops take significantly longer to perform as a family than one thirty-something man requires at a highway rest stop to get moving again. Our first few long car trips as a family I was frustrated because we were spending at least 50% longer "on the road" when visiting family than we ever had before. My wife was frustrated because for her all these delays and stops were already baked in to her mental estimate and she couldn't understand why I wasn't already taking them into account.

This situation didn't cause any individual catastrophe or project failure, but there were a few times that I mentioned a potential arrival time to my parents that was off by more than a few hours. That caused some irritation when plans had to be adjusted to account for the real travel time. A better (more thoughtful) estimate on my part could have prevented this.

It took me months to recalibrate my travel time estimation hunches to account for the new reality of parenthood. While I trained my brain to think of trips as longer I used my wife's ballpark system of just tacking on 90 minutes to whatever the navigation app would say when we got into the car. It seemed unreasonable to add so much time to our estimates but again and again her hunch proved correct and I started to adjust my internal conception of how long a drive with kids takes.

This anecdote might seem like it doesn't have anything to do with IT Project Management, but I've seen the same pattern manifest with people who try to ballpark estimates using intituation and experience that was gained from previous jobs or situations that's no longer applicible to the situation on the ground.

As a software developer early in my career I would often see managers unfamiliar with the transition from on-premises servers to cloud-based development. Their experience with servers was a clunky process that would involve sourcing physical product, taking delivery, configuring local networks, installing physical media, and deploying an operational unit over the course of weeks.

There was often a conceptual disconnect when system administrators on a project would talk about new server deployments in terms of hours instead of the days that management expected. Thankfully it was more often than not in favor of a faster project but definitely taught me that while hunches and intuition are important for estimates they constantly need to be checked against the actual facts on the ground.