Here's a basic queuing theory fact that may have behavioral/psychological implications.
If customers take on avg 10 minutes to serve and arrive randomly at a rate of 5.8 per hour, then with one bank teller working, expected wait is 5 hours. With two tellers, 3 minutes.
1/
If customers take on avg 10 minutes to serve and arrive randomly at a rate of 5.8 per hour, then with one bank teller working, expected wait is 5 hours. With two tellers, 3 minutes.
1/
I learned this from johndcook.com/blog/2008/10/2…
The reason for the counterintuitve difference is that 5.8/hour means customers arrive on average 10.3 min apart, so there's little slack. If they arrived equispaced, everything would be fine. W/ random arrivals, pileups are awful.
2/
The reason for the counterintuitve difference is that 5.8/hour means customers arrive on average 10.3 min apart, so there's little slack. If they arrived equispaced, everything would be fine. W/ random arrivals, pileups are awful.
2/
Two speculative implications for the self-managed knowledge worker. First, if you accept tasks subject to the rule that inflow < outflow (so that your queue does not grow unboudedly), BUT you make inflow nearly EQUAL to outflow, you'll sometimes have huge delays.
3/
3/
The counterintuitive example works because the system behaves very badly when the expected time between arrivals is nearly equal to the completion time. If you stay away from that regime (leave 20% of time as planned slack), your delays/stress go down by WAY more than 20%.
4/
4/
My second implication is more speculative, but maybe more interesting, and I'll phase it as a question.
Can patterns of self-management change whether a single knowledge worker looks more like one or two bank tellers?
5/
Can patterns of self-management change whether a single knowledge worker looks more like one or two bank tellers?
5/
The problem with one bank teller is that it has to wait until a slow, problem customer is handled before calling "next."
That reminds me of how I feel when one hard or unmotivating task is getting done very slowly, blocking normally-arriving new tasks.
6/
That reminds me of how I feel when one hard or unmotivating task is getting done very slowly, blocking normally-arriving new tasks.
6/
It seems to me that the kind of work practices advocated by Terry Tao here alter the queueing regime to look more like a multi-teller one, with several tracks where some things are guaranteed to be accomplished.
terrytao.wordpress.com/2008/08/07/on-…
7/7
terrytao.wordpress.com/2008/08/07/on-…
7/7
PS/ Big thanks to @sogawasan for the pointer that led to this!
• • •
Missing some Tweet in this thread? You can try to
force a refresh
