I think indirect expenses per employee would probably be a much better metric. Amazon buys and sells lots of stuff, which pushes up both their revenue per employee and their expenses per employee; similarly, Google pays websites to place adsense adverts and then resells the space via adwords. On the other hand, eBay just puts buyers and sellers together and doesn't act as intermediary, so the value of goods sold via eBay isn't reflected in eBay's revenue or expenses.
It's a basic maxim of economics (and accounting) that equivalent transactions should be treated equivalently; and whether goods pass directly from seller to buyer (as with eBay) or through an intermediary (as with Google or Amazon) therefore shouldn't affect how we compare the companies.
It's not the appropriate number. Profit margin makes more sense.
ROI is useful in a case where you make an expenditure now, and get back money (or savings) in the future. For example, you automate shipping for $1 million, and save $300,000 a year, for an ROI of 30%.
It doesn't work for employees, except to the extent that you're taking an up-front hit ("We paid for her to get a Master's degree") in exchange for something else ("Now she's more productive").
Managerial accounting was my area of specialty; I was simply pointing out that there is also a cost associated with treating employees well. I think Google's numbers may be a bit high because it has so many people who work for it for free.
Now, if we could get some metrics about salaries, this could become even more interesting.
Why is that more interesting? It looks like you're dividing revenue by employee count. A high figure there just gives you efficiency, effectively the same thing that the profit per employee gives you. It says nothing about how expensive the employees are, Google isn't actually paying 900K per head.
It looks like you're dividing revenue by employee count
No, not dividing, only subtracting. Accounting 101: Income - Expense = Profit. The article gave us two pieces of the data: Income and Profit. If something is neither income nor profit, it is an expense.
Grinding down the details into something like "profit per employee" tells only half of the story; it's like looking at only one side of a balance sheet and basing conclusions on that.
It says nothing about how expensive the employees are
I wasn't implying that any of the top 3 were actually paying 899K or 871K or 863K per head, but if that is the amount of resource expenditure it requires to make X amount of profit, it certainly does reveal a bit about how companies are managing their resources . . . a metric any investor would be interested in knowing.
Again, it goes back to efficiencies. Google has few employees for the amount of revenue it generates (high revenue per employee) but also has fairly high costs (millions of servers). They could be paying all of these people minimum wage and your formula would still show high expense per employee.
I get what you're trying to show, but calling the employees expensive because there happen to be a lot of expenses is misleading. They are only expensive if the expenses actually related to the employees.
I think this accounting discussion at the per-employee metric is getting out of hand. We're facing a very real apples-and-oranges conversation on expense when comparing Oracle (software) versus Amazon (retail). The margin -- and revenue -- focuses are completely different. If Amazon sustains growth with a 15% profit margin, that's cause for celebration; if Oracle's margin dropped to 15%, heads would roll.
http://zentu.net/snaps/exp.png
(Disclaimer: I studied accounting)
Amazon, Google and Apple also have the most expensive employees; the cost-per-head is almost 900K.