I have a team which works on online jobs; their productivity is measured in terms of the number of jobs completed per hour. The team has been given a goal for this metric. Let's call the task as Task X and the metric Tasks Per Hour (TPH).
The jobs are assigned to each team member through an online queueing system. The team is pretty large (~800). Their primary task is something else (Task Y) and the Task X is secondary for them. The secondary task (Task X) is less than 10% of the main task (Task Y). The queueing system sends both Task X and Task Y randomly to each team member. Due to the large size of the team, the number of secondary tasks assigned to each team member is typically small (4-5 per shift); however, creating a smaller team to handle this secondary task is not feasible due to certain business restrictions.
Recently, the team's TPH for Task X (the secondary task) has consistently been below the goal set for them. I have to find answers to the following:
a) Is the current productivity goal correct or is there a need to change it?
b) If there is a need to relax the productivity goal, what should it be?
I am looking for guidance on the statistical tests / tools which can be employed to undertake these tests. I have access to the performance of the team in both primary and secondary tasks.
Thanks in advance!
Hello George,
There may not be a single, correct answer to either of your questions. But, I suspect you'd have more insight if you tried something along these lines:
1. Have supervisor(s) rate a representative sample of incumbent workers (preferably not brand-new ones, as the ability to juggle and complete tasks efficiently may require an on-the-job learning curve). This rating should be based on information (e.g., direct observation, peer recommendation, success of handling Task Y assignments, other sources) other than the "TPH" scores on Task X.
2. Sort the list from highest to lowest rated.
3. Determine the actuarial/observed outcomes (TPH on task X) for each of these.
4. If it appears that there is a functional relationship between rating and TPH, then:
5. Select a rate that would not exclude workers judged by supervisors to be satisfactory or better in their job performance. This is not unlike a sensitivity/specificity analysis (though, you're playing with peoples' livelihoods).
If there is no functional relationship (step 4 above), then select a rate that is met by at least 85% of workers (or whatever amount of workers you don't want to frustrate unnecessarily).
Note that a host of factors can muddy these waters quite quickly, such as: (a) the actual mix of X/Y tasks assigned; (b) the complexity of a given task; (c) whether tasks are closed immediately or require partial completion, then return (juggling plates in the air, like a circus performer); (d) length of job tenure, etc.
Good luck with your work.
- Badges
- Science topic
- Similar topics
- Biological Science
- Physiology
- Physiological Processes
- Stress