I'm wondering if there is a good theoretical reason why nobody seems to use Marginal Analysis to optimize how much human talent is stocked in inventory. In all the classical treatments in textbooks, newspapers, donuts, bread etc have been analyzed using Marginal Analysis. It's always a widget that sells or doesn't sell within one inventory period.
Well, check this out: I know of specific firms where they contract with human talent, and sell that human talent to customers. Examples of talent for which there is a limited weekly supply (for example) and a definite demand, include former US presidents who give talks; tech support on the phone; management and technical consultants; astrology / entertainment. You might notice that a specific talented individual may be limited to a certain supply of hours they can provide in a given week. There is also some sort of demand for the talent to purchase their services.
How many hours would a smart firm want to stock, ie., purchase from such an individual, to book their time in advance? How many hours per week should the firm contract with the talent? This is the essential question.
Inputs to the calculation:
Given ML and MP we now can compute P statistic = ML / (ML + MP). Now we can use probability distributions (discrete or continuous) to find the optimum "stocking" quantity for the talent. In this case it will be number of hours "on" that is contracted in one week for example.
Is Marginal Analysis suitable, or is it unsuitable for optimal stocking quantity of saleable human services? If not, why not?
Thanks for reading.
asked 12 Feb '13, 12:09
On the surface, what you are describing is the classic "newsvendor" problem, only with multiple commodities. You could model the contractual arrangement with each individual (e.g., speakers) or group of commensurable individuals (e.g., level one tech support specialists) as a newsvendor problem. There are some difficulties, though. One is the issue of substitution: excess inventory for one individual/category may be useful to satisfy demand for another individual/category, but perhaps at a different revenue value (and possibly with a different cost -- I charge more for things I'd rather not do). Another is that, unlike newspapers, some individuals can provide multiple distinct services.
So those two wrinkles may account for a lack of inventory modeling of human assets. Another possibility is that it is in fact being done, but is not being published (or discussed) because it represents a proprietary method that provides competitive advantage to the user.
A third possibility (one that I think is a significant contributor) is that the people who manage human resources tend not to be the most quantitative by nature. I think that college HR majors rarely require any math beyond the general college requirement, and perhaps some statistics. When I taught quantitative modeling to MBA students, the HR concentrators (with one notable exception) tended not to dominate the top quantiles.
answered 12 Feb '13, 14:20
Paul Rubin ♦