The Software Industry's Productivity Boom: Analyzing Revenue per Employee Trends
Recently, I was on the HR Heretics podcast and we talked about the increasing efficiency of software companies (in addition to other topics including the implications of AI for executives, how to diligence a candidate, & what board members expect of their people leaders).
I realized it has been a decade since I’ve updated revenue-per-employee metrics.
Here are the trends across the group of 12 publicly traded software & infrastructure companies (which are the fastest growers or most-highly valued) over the last 5 quarters.
Revenue per employee spans approximately $100k-$400k. Salesforce tops the list. ServiceNow, Workday, & Veeva round out the top 4.
Compare that to the benchmarks in 2013! In addition to the dated charts!
In 2013, the average revenue per employee of these companies totaled $200k. Today, that number is $470k for the basket of companies above, a 135% improvement.
Normalizing for inflation (which was 34% during this past decade), modern businesses produce 2x more RPE than their decade old selves.
|Company||RPE 2013, $k||RPE 2023, $k||CAGR|
Most companies in the image have been acquired. The four horses of the previous decade that still run free, I’ve listed above. RPE for the group grew from 1.8% to 9.6% per year.
In 2023, these companies added about $21k in revenue per employee, but the range spans $13-29k.
The reality is most companies do not grow revenue-per-employee per year linearly. Here’s a rough estimate of RPE over time.
|Year in the Company’s Life||Estimated RPE Growth||Note|
|15 +||Late-stage public||1-5%|
For these publicly traded companies, there’s a slow decline in growth rate but the combination of product-market fit & the opportunity to expand the product portfolio allows them to keep expanding revenue per employee at attractive rates.
I wonder what trends we’ll see in 2033. Perhaps another doubling (driven by AI!)
- Note this update post reflects accurate data. The first version of this post published on 2023-11-13 used data from an inaccurate data source. I regret the error.