“It's not that we have little time, but more that we waste a good deal of it” – Seneca
In one of today's training sessions through ProcessMiningIQ, I stumbled across a good lesson on time. Or, rather, how you measure it. As it turns out, Celonis may be causing you to underestimate the duration of your process.
Celonis' PQL has a few quirks about time. First, take a look at the following table:
The first column is the case key. The second is the default KPI you can select for Total throughput time in days. This calculates the days between the first and last activity of the case using the function CALC_THROUGHPUT. You'll note that this is rounded to the nearest one's place despite the configuration being for two decimal points.
The third column is the days between the first and last activity of the case using the function DAYS_BETWEEN. Here you'll see that we do have all the way to two decimal points. So, what can we take away from this? Well, by defaulting to using the standard KPI from Celonis, you'll be undercounting the number of days. Why is this? Because Celonis always rounds down! Regardless of the decimals you see in Days Between, the value in the second column will always be rounded down. This could lead to a costly miscalculation if not taken into account.
Finally, let's move to the third column, Months Between. Having done the math, you'll see that Celonis is taking the time between the two data points and dividing it by the average length of a month. It seems therefore, that this is more of a comparison metric and not actually returning the literal value of months between the dates.
Hope this was a timely update for you as you continue to learn about Celonis and Process Mining!
Komentar