Latest R&D Metrics Show How Bad Products from Payments Companies Likely to Be

We’re starting this article with a chart. But we’ve removed the title and Y axis of the chart so we don’t bias anything at first blush. Ready?

This chart is clearly up and to the right, is it not? From the first quarter of 1947 to the first quarter of 2012 it looks like the value of the line has gone up 5x. If this line was indicating something positive – like money in your bank account – you’d be happy that it’s up and to the right.

What this line is showing is labor productivity. We sourced it from the Bureau of Labor Statistics – below is the chart in its entirety.

If you’re wondering what labor productivity is we’ll copy some text from the BLS for you, complete with a useful example.

Workers in the U.S. business sector worked virtually the same number of hours in 2013 as they had in 1998—approximately 194 billion labor hours. What this means is that there was ultimately no growth at all in the number of hours worked over this 15-year period, despite the fact that the U.S population gained over 40 million people during that time, and despite the fact that there were thousands of new businesses established during that time.

And given this lack of growth in labor hours, it is perhaps even more striking that American businesses still managed to produce 42 percent—or $3.5 trillion—more output in 2013 than they had in 1998, even after adjusting for inflation. One might wonder how such a large amount of additional output came into existence, given that American workers did not put in any more hours of work in this most recent year than they had 15 years earlier. One thing can be said for certain: the entirety of this additional output growth must have come from productive sources other than the number of labor hours.

Labor productivity is defined as real output per labor hour, and growth in labor productivity is measured as the change in this ratio over time. Labor productivity growth is what enables workers to produce more goods and services than they otherwise could for a given number of work hours. As an example, suppose workers in a factory can make 20 cars an hour. One month, the company modernizes machinery and the workers take training classes to help improve their performance. Using the new machinery and recently acquired knowledge, the same workers can now make 30 cars an hour—which is a productivity gain of 10 cars per hour, a 50 percent gain

https://www.bls.gov/opub/btn/volume-3/what-can-labor-productivity-tell-us-about-the-us-economy.htm

Labor productivity is how much of a good or service can be produced with an hour of labor. If the labor productivity number goes up – as it certainly has – it’s because the labor is getting more efficient. Thanks to things like software and automation an hour of human labor can produce much more today than it could 50 years ago.

Another way to say this is that the cost to produce a good or service had dropped substantially over time. Here’s some math for an example.

A company makes sweaters and sells them for $5 each. Each employee can make 10 sweaters a year. The company hires 10 employees to make 100 sweaters and then sells them all for a total of $500. The cost of each employee is $25 per year. The profit for this sweater business would then be $500 – 10x $25 = $250.

The company invests $50 in automation. Now one person can make 100 sweaters a year. The cost of the employee is still $25 per year and the sweaters still sell for $5. Even without growing the business revenues, which are still $500, the company profit has gone way up to $425.

Here are the revenues per employee for IBM.

Now here they are for Facebook, Google, and Amazon, the latter of which is seeing their metrics drop as they hire relatively low wage factory workers.

Excluding Amzaon, up and to the right, correct?

Now, let’s look at something else. Would it surprise you to learn that the companies getting the most growth in labor output also happen to be the ones that are investing the most in R&D, or general software engineering? Call us crazy, but software is sort of like this force multiplier for automation.

From our venture capitalist friend Sammy Abdullah,

The general rule of thumb for spending in SaaS is 40/40/20. In other words, 40% of operating expense should be on R&D, 40% should be on sales and marketing, and 20% should be on G&A. 19 SaaS companies have gone public in 2018 and 2019, so we wanted to see what their ratios are. 

On median and average, 29% of OPEX is R&D, 48% is on sales and marketing, and 22% is on G&A. Overall, it’s pretty close to the 40/40/20 rule, but really 30/50/20 may be more accurate. 

Sammy Abdullah, Blossom Ventures

Sammy’s data is below, and he even has an updated article to convert these numbers to investments as a percentage of revenues (median SaaS values are 24%).

So the companies getting growth and positive labor outputs are investing 30% of OPEX into R&D? What a concept.

Now let’s revisit an earlier article analyzing our friends in the payments processing ecosystem. Here’s how much these guys spend on R&D by comparison.

• EVO Payments: spent $44.3M on “computer software” relative to 2018 revenues of $565M, or 7.8% of revenue was spent on R&D.
• Fiserv: spent $496M on R&D relative to 2018 revenues of $5.82B, or 8.5% of revenue was spent on R&D.
• Global Payments: spent $213M on R&D relative to 2018 revenues of $3.366B, or 6.3% of revenue was spent on R&D.
• First Data: spent $143M on R&D relative to 2018 revenues of $9.498B, or 1.5% of revenue was spent on R&D.
• NCR: spent $252M on R&D relative to 2018 revenues of $6.41B, or 3.9% of revenue spent was on R&D.
• Square: spent $497M on R&D relative to 2018 revenues of $3.3B, or 15% of revenue was spent on R&D.
• TSYS – spent $39M on “Additions to internally developed computer software” relative to 2018 revenues of $3.816B, or 1% of revenue was spent on R&D.
• Worldpay: nominally spent $1.07B on R&D (it includes plant, property, and other capitalized items so we expect this number is grossly inflated) relative to 2018 revenues $3.925B, or $27% of revenue was spent on R&D.

We should also note that some of these R&D figures are grossly inflated due to the payments entities acquiring software companies. For example, in 2017 Global Payments bought ACTIVE networks, and in 2018 it bought AdvancedMD. Here’s the embarrassing amount Global Payments invested in software in 2016 prior to these acquisitions per their 10K:

We made capital expenditures of $91.6 million and $92.6 million during fiscal 2016 and 2015, respectively, including $36.5 million and $35.1 million of internal-use capitalized software development costs.

GPN 10K, 2016

$35M on revenues of $2.9B, or a paltry 1.2%. And we’d bet that Global’s management intends to find ways to slash R&D out of their recent software acquisitions, too.

Now let’s look at a table of public companies with the fastest shrinking revenue per employee metrics. Look at the companies we outlined in red – they’re payments companies!

Coincidence? Of course not.

But the most fascinating component here is this:

Even though the entire US economy has been drastically increasing labor output, and even though payments companies by and large are neglecting even reasonable levels of investment in R&D, they’re still increasing the fees they charge on their undifferentiated and unimproved product.

This is just perverse. The products they hock – payments processing – should cost much less to produce today than they did even a decade ago. And it’s not as if the payments companies are investing money in anything, so why do they keep adding make-believe fees to increase the prices of their processing?

There’s no other way to look at this except that the payments companies are either massively incompetent and have no idea how to use automation and software that every other company seems to be benefitting from, or they’re outright thieves.

How this industry hasn’t been more closely legislated is patently insane.