Saturday, January 30, 2010

"Loading" and "Productivity"

After measuring productivity quarter over quarter, for a large pool of young and skilled engineers, I see a pattern which may not be too obvious.

It was known that if the "loading" is less, productivity will be less.
Meaning, say you have 5 things to get done with 1 person's help.

We assign 1 item out of the list to the person. Wail until that is done to assign the next one etc. The productivity will be lower than assigning 2 from the task list together.

Logic: Person has to do certain level of multitasking. When stuck on one task because of external dependencies can work on moving the 2nd item. etc.

Now, when we stretch this, assign all 5 tasks you know of for the person at once without serializing or prioritizing them, or telling each one is critical.

The productivity of this model becomes lowest.

This one is more tricky ! Here the reason of losing efficiency is no longer because of "idle cycles because of waiting on external dependencies". It is more behavioral !

Human being tends to "procrastinate" when there is "too much" to do !

Right when I think I am getting to the right level of loading for optimal efficiency of the work force, my trainer at the gym broke the news to me - "If you overwork, you ll feel so sour that you wont come to gym for 3 days, and that will make you achieve less result. If you work out too mildly, your body would stay lazy and feel no need to build more muscles."

Hmmm.. this is not new !

No comments: