The Labor Floor
Why and when it doesn't make sense to automate - and what kind of jobs are safe?
Automation is not new. Newspaper clippings from the ‘60s and the ‘50s echo many of the same stark warnings about automation as we hear today. This article from POLITICO details an era where an ongoing economic recession was stubborn, and many people at the time thought that they were in an “Automation Depression,” an era of machines slowly replacing workers.
“A part of the current unemployment … is due to the automation component of the capital-goods’ boom which preceded the recession. The boom gave work while it lasted, but the improved machinery requires fewer man-hours per unit of output.”
Well, we came out of that recession in 1958, and the economy is booming, and unemployment is at a historic low. Automation is a hot topic again, and this time with headlines focusing on LLMs and humanoid robotics. I’m here to tell you that it’ll be a while until all jobs are automated, and this is due to what I call the labor floor.
Calculating the Labor Floor
Automation, at its core, is a capital expenditure. You pay a high one-time implementation fee (or development cost) and usually a low overhead to replace all or a part of a labor workflow in anticipation of an ROI.
An example labor workflow is below:
Unload truck
Unpack boxes
Organize in the appropriate warehouse section
Log order
Fulfill order
The primary goal is to save on labor costs. There are, however, ancillary benefits such as error reduction or productivity/throughput gains. These are tougher to forecast but are key in crafting the entire business case to implement an automated solution.
Let’s focus on the labor cost reduction part, which is usually the easiest to calculate. Taking the steps above, I’ve crafted a simple labor matrix to show the costs associated with each step.
In this matrix, we see the steps, the labor hours needed for each step, the hourly cost of labor, the cost to automate (this can be variable due to the complexity of the automation solution), and the hourly labor cost after automation.
Now, after calculating the yearly cost & payback periods, the most labor-intensive step (#1) makes the most sense to automate. The payback period length of #2, since it’s a harder technical problem with lower labor hours, is way longer than #1, indicating that it doesn't make sense to automate that step for now.
Take a look at step 5, a technically non-difficult automation step; hence, the lower cost to automate. It makes sense to automate this step and can provide a payback period of just over 4.5 years even though it is the same labor cost as #2.
Automating end-to-end workflows is tough, and when you break down the workflow into steps, sometimes certain steps just don’t make economic sense. (I wrote a bit about how Tesla failed to navigate this problem correctly in this past article.)
Your job is probably safe - for now.
Herein lies why I think many jobs are safe for now: automation is hard to customize. Humans are incredible at adapting, and hardware or software automation development comes at an increased cost the more you need to customize it to a specific workflow.
For some jobs with core aspects that automation is not great at, such as adaptability, variance in task profiles, or constrained/ridged environments, the capital expenditure to develop a custom solution does not justify the labor savings, even when you can forecast better productivity or fewer errors. This is especially true for jobs with a low labor cost, where the market is currently dictating that it’s simply more cost-effective to retain human labor for now.
Even if large firms come in to help finance the CapEx, plotting savings on a 5+ year ROI is risky, as technological advancements in this fast moving space can leave firms with defunct technology.
The point of all of this is that many jobs will be safe, until the market brings cheaper technology & cost of implementation or labor rates rise where the payback period for a new automation solution can be justified.
This doesn’t mean you should feel safe; technological breakthroughs happen constantly. You should get prepared for an automated future and subscribe to Prompter to understand how to stay ahead of the curve.
Stay ahead of the curve and get comfortable with automation and AI today. See you in 2 weeks!
— The Prompter Team