Software eating the Fortune 500

I've been thinking again about the software eating the world thesis. Recently, I came across two related things.

First, this Fresh Air interview with Steve Coll on his new book about Exxon Mobil. He commented along the line that in 1955 Exxon was #2 on the Fortune 500 list. In 2012 they are #1. In another 50 years, would you expect Apple or Exxon to still be near the top of the list?

Second, Mary Meeker's slides from the All Things D conference on the state of the Web. Slides 32-84 show examples of the re-imagination of X industry/product, which is the most visual manifestation of the thesis I've seen to date.
One interesting property of software is that it is much easier to make tweaks to user interfaces. As software eats the world, people are interacting more with software in all aspects of their lives.

As a consequence, we should expect people to interact with things that are changing more rapidly than before. It's human nature to adjust to surroundings, so I suspect over time people are just getting used to things changing more rapidly.


At the same time, the Internet has enabled software to spread at unprecedented rates. Companies are getting scale faster and faster. 

OK, so where am I going with all of this?  It seems that a corollary to the software eating the world thesis is that we should eventually be seeing bigger companies form and rise faster and faster over time. Not only should they pick off huge incumbents, but they should also pick off each other in faster waves.

To examine this corollary quickly, Hunter Lang and I did some quick crunching of Fortune 500 data. I'm sure other people have studied this stuff much more deeply, but I couldn't find anything quickly and I wanted to mess around a bit myself. Here's what we found:


We took each year from 1955 onward and then counted how many companies were present in the previous year (in blue) and five years before (yellow).[1][2]

The missing data was due to anomalies in company names in the data. There were other anomalies, but this is the only one that seemed to cause a major shift in the trend.[3]

There are probably much better leading indicators of this phenomenon, but the trend seems already clear that things are turning over quicker over time.[4]

We might predict that we'd see a continued downward trend and at an increasing rate. Yet there seems at least several things that could counteract this force, at least for the Fortune 500.

1) The bottom company in the Fortune 500 is still a huge company: 4.3B in revenue. There's a long lag time to get to that level, so we may need to wait a while for the current crop of startups mentioned in Meeker's slides to have any meaningful impact on this graph.

2) The Fortune 500 seems based on revenue, which can make software companies, which generally have relatively good margins, look worse.

3) It's very possible we may see increased movement quickly in a wider universe of companies (say top 3000), but the Fortune 500 might remain more constant because they acquire the up and comers. Opsware comes to mind as anecdotal evidence, which I choose because it is coming at the horse's mouth of the software eating the world thesis. Of course the IPO market has a lot to do with this area.

4) As our government regulatory system expands via the seemingly inexorable growth of administrative agencies, we would expect more and more regulatory capture.

5) I didn't look too hard, but the upticks in the graph seem related to economic cycles. If we go through some major ones, that could do all sorts of things, but probably just delay the inevitable. That is, economic cycles don't stop the growth of technology and there is evidence that the eventual down-tick is a result of companies often formed in recessionary times.

It's an interesting thought experiment to think what the floor is on this number. One day soon will only 250 companies in the Fortune 500 have been there 5yr ago? 200?

[1] We decided to count in/out changes instead of ranking changes so we didn't have to deal with the infinity problem of people dropping off the charts. Also ranking changes just seemed to enter more noise.

[2] The hypothesis was a 5-yr thing because that seemed like enough time to let faster moving industry dynamics take shape, but not too much time to reduce our data points.

[3] The counts of unique companies in the data were sometimes less than 500 per year as the same company name would appear twice in the list -- not sure why. We also didn't look too hard into the major anomaly, so it very well could have more pernicious effects on our graph than we think.

[4] I included the 1yr line as more of a sanity check, but really it may not be much of one if they change/normalize the names of the (same) companies around every 2-5yrs. However, even the 1yr line shows a definite downward trend.


If you have comments, hit me up on Twitter.
I'm the Founder & CEO of DuckDuckGo, the search engine that doesn't track you. I'm also the co-author of Traction, the book that helps you get customer growth. More about me.