The WSJ ran an article on the effect of COVID on small businesses. Some of the information is shocking — from the difficulty of getting capital (even through government subsidized loans), to the fact that perhaps as many as 1 in 5 businesses that were open in January may have shut their doors permanently. But actually, while that is an extremely painful number, the truth of the matter is, it’s much better to save the entrepreneur than to save the business for a few more months and have both it and the entrepreneur fail. The entrepreneur can start again — as long as he or she hasn’t been crippled. But the more interesting thing for me was a graph they presented, that looked something like this:
Their point was that, for the first time, a larger percentage of the US was employed by giant businesses, over 5,000 employees, than by small businesses. But when I saw that, I was terrified. My first thought, looking at those curves and the dates, was, “wait! Has the Internet killed small businesses?” But then I thought about it some more, and started digging into the data, and realized they were engaged in some funny business. They’ve changed the scale of the chart, and they’re comparing a particular set of numbers. Here’s another way to look at the exact same data, with more granularity on company size, and with everything adding up to 100%:
A few things to note: 1) while the 5,000+ category has grown, it’s not nearly as scary as it looks in the WSJ chart; 2) there’s no giant change around 1995 indicating some Internet phenomenon, rather just the curve continuing its trajectory from at least 1988; and 3) actually, most small businesses seem to be employing about the same number of people — it’s actually the 20–99 number that’s shrunk, and a fair bit of that has gone into the 100–499 bucket and the 500–4,999 bucket. There are a lot more questions that one can ask here. How much of that change is small businesses growing into big businesses? How many businesses does this actually represent? I could go on … and I probably will, but I’ll have to dig into the numbers more, so we’ll save that for next week. Bottom line, this is an example of the comment I made above about even stolid publications (like the WSJ) engaging in nonsense — the chart they showed was borderline deceptive.