Estimate Activity Durations

With excerpts from my Project Management Novel, I will illustrate the many processes of the PMBOK.  Here is the ninth one: Estimate Activity Durations. Use this map to see how this process fits into the scheme of processes.  Estimate Activity Durations

"Now, how long do these activities take?”

The men gave their durations and Gwilym wrote these numbers on the activities.

The team continued in this way, placing and rearranging activities on the bull hide and writing durations on the activities. The intricate structure of the work revealed itself to the team. Some activities were added during this session and other activities were determined to be unnecessary when looked at in the light of the overall project. When this happened, Fred made the appropriate adjustments to the Work Breakdown Structure.

At some points, the men took a while deciding on the duration of an activity. The first time Gwilym saw a man struggling, he asked “What’s going on inside your head right now?”

“Well,” replied the man. “I’m thinking about the last time I did this job and how long that took. It was eight days. But this job is about twice as big so perhaps it’s sixteen. But the weather was horrible last time and this time it looks like we’ll be doing it in June, not December. So I have to make adjustments for that. So I’ll say twelve days.”

Gwilym broke out into a broad smile and turned to the rest of the men. “See what he did there? That’s exactly what I want you to do. If you’ve done the activity before, and remember how long it took, use that as an example for how long this new activity will take. And don’t forget to make adjustments like Frank here did based on the realities. That is called…” he turned to Fred, “Analogous Estimating.”

When the time came to decide how long it took to build the main walls of the tower, the head carpenter talked out loud. “Each log will take an hour of carving to get the notches just right but we can do that with one group of men while the other places the previous log. So the time to do this activity is the time it takes to place each log properly in the tower. That will be pretty quick for the first few logs but will take longer as we have to start using winches to put them in place. So the ground level logs will take one hour apiece and the higher logs will take about two hours a piece so that we’re careful not to drop them.”

“Good,” said Gwilym. “How many hours does that make it all together?”

The man calculated in his head. “Four logs per level, four high before we need to start using the winch, so it’s sixteen hours for the first eight feet. That’s two days with expected problems and setting up the winch for day three. Then it’s two hours per log after that, that’s one day per two foot layer. That’s sixteen more days. So a total of eighteen days to raise the logs to the top.”

Fred asked, “That’s different than ‘Analogous Estimating’ right? What do tha call it?”

“He’s using parameters and multiplying them by the number of units. Let’s call it “Parametric Estimating.”

Later the carpenters were arguing about the duration for building the stairs. As usual, Gwilym asked them to state the assumptions they were using to get to their estimates. One was assuming warm weather, the other was assuming cold and rainy weather. Since the stairs looked like they would be built in the early spring, it was unclear which assumption was most likely. Gwilym suggested a compromise. “Let’s assume first that the weather is perfect the entire time. How long would it take to build the stairs?” The two arguing men agreed on twelve days.

“Now we’ll assume it is cold and rainy the entire time, maybe even snow. Then how long will it take?” The men talked amongst themselves for a while and came up with a duration of forty days.

“Now we’ll take the most likely estimate. Assume a typical spring, with some nice days, some rainy, one day of snow. Now how long will it take?” The two men agreed on sixteen days.

“Good!” said Gwilym, writing these three numbers on the wall. “Now we do some mathematics. We take the optimistic estimate, add to it the pessimistic estimate and four times the most likely estimate and divide the result by six. So we have 12, plus 40 plus 4 times 16 equals 116. Divide that by 6 and we get a little over 19. So let’s estimate 20 days for this activity.”

Duration estimate = (T0 +4xTr +Tp)/6  = (12+4x16 + 40)/6 = 116/6 = 19.3days

The men appeared impressed. Fred asked him, “What do tha call that one, Gwilym?”

“Three-point estimating,” he replied. “They use that technique for calculating caravan journey times in the east.”