Compound annual growth rate (CAGR) applied to non-financial growth I’m trying to understand if “compound annual growth rate” is appropriate when applied to non-financial measurements.  For example, when that term (and the appropriate calculation) is applied to money growth, it makes sense because the money we earned in the previous period is added to the principal and is contributing to the growth of our money in the current period – i.e., the compounding effect. 
However, I feel as though the term (and the calculation) is inappropriately applied to many non-financial measurements. For example, let’s say we’re measuring “units sold” instead of money.  Since the units we sold last period do not contribute to any growth in units sold in the current period, it doesn’t make sense to me that we could use the same term (and calculation) to interpret the results. 
While there are lots of references across the ‘net that use CAGR in non-financial calculations, they all seem to reference sources that trace back to examples using only financial measurements.  Calculating population growth would be one of few non-financial measurements where CAGR makes sense, so do most people get it wrong by using CAGR for non-financial calculations such as units sold or is there something I’m misunderstanding?
 A: 
Since the units we sold last period do not contribute to any growth in
  units sold in the current period, it doesn’t make sense to me that we
  could use the same term (and calculation) to interpret the results.

Consider radioactive decay. It's described by the following equation: 
$$N=N_0e^{-\lambda t}$$
This is essentially the same equation as a CAGR, e.g. in discrete compounding  $CAGR=e^{-\lambda}-1$. 
The mean lifetime is a familiar $\frac{1}{\lambda}$, which you recognize from weighted average life (WAL) equation in finance.
The essential observation here is that the process is exponential growth. We often model sales as exponential growth too. There are a couple of reasons to do this. One is that sales are in dollars and CPI will make them grow at the rate of inflation. The second is that the sales grow because of increasing customer base, marketing and capital expenditures. Here's an example of Google Play revenue:

This looks exponential to me. Your argument about past sales not influencing future is weak. Because as I wrote earlier past sales indicate the customer base and the past capital expenditures. Also, the fact that you bought something in Google Play could mean that you may buy again there. That you purchased is probably not the last one you'll ever purchase. And the mere fact that you bought something means that you are now a customer, know how to buy stuff on the store etc.
