Computing growth rates with drastically changing sample sizes

Question

I'm computing growth rate for some companies annually for 3 years using this formula, for each year:

((# customers year 2 - # customers year 1) / # customers year 1) * 100

The problem I'm running into, though, is that the # of customers is changing drastically for some companies...from starting a business, going from quite small to decent size. So the growth rates look inflated and I'm wondering if there is some other way I should be calculating Growth Rate for this situation.

Here's an example of what I'm seeing:

year # customers
2014 1

2015 40

2016 100

2017 130

Growth Rate 2014-2015 = (40 - 1) / 1 * 100 = 3900%

Growth Rate 2015-2016 = (100 - 40) / 40 * 100 = 150%

Growth Rate 2016-2017 = (130 - 100) / 100 * 100 = 30%

Answer 1

That's the growth rate. You can use log differences, but that won't change much.

So the usual way to deal with this in corporate finance is to introduce the scale/size factor in your analysis. It can have many different forms, e.g. adding the size variable in regression models, grouping firms by size when comparing etc.

The growth rate itself - as opposed to the absolute growth delta - is supposed to deal with different sizes, but it only works when the growth rate itself is independent of the size, which is not always the case.