Why Training Workers Costs More Than You Think
I read Adam Davidson's article in the Times Magazine last week with interest. In it, he discusses the difficulty that one Milwaukee employer has had finding workers, with particular attention to the problem of the costs of training workers in high-tech skills. In a nutshell, he talks to an employer who is offering a very low starting salary ($10 an hour) and then complains that he can't find applicants with the skill set he requires.
I'm not very sympathetic to these complaints. It's been my experience that $10 an hour doesn't get you much in the labor market. Maybe that's a reflection of the higher cost of living around Philadelphia, but I wouldn't dream of hiring someone for less than $12 an hour. Pay any less than that and your worker will not be able to afford a reliable car and probably won't have a good housing situation, either. That adds up to trouble.
My former partner told me that 90 percent of human resources problems are generated by the lowest-paid workers in the company, and I have found that to be true. I can't speak for anyone else, but I have found that people I'm willing to pay $15 an hour are generally pretty reliable, and those who are worth $20 or more are solid, dependable workers who can be expected to perform at a high level and acquire significant skills. That's a level of pay that also allows them to have stable personal lives. They can afford a car, and, with a working spouse, a house and children.
I benefit from having settled workers. People who show up reliably, work hard and realize that they benefit when the company prospers. When I take the time and go to the expense of training people like this, I can have a reasonable expectation that the investment will pay off for all involved. Training is not cheap, though.
Here's an example of what training costs me. We are at the point where having a single design engineer is not enough. Our sales are growing, and we now have periods when he can't keep up. I don't need another engineer all the time, just some additional capacity to handle the busy periods. Eventually, if sales keep growing, the second engineer will be a full-time position. But not for the next year or so.
The classic solution to this is to outsource the engineering until we have enough demand to bring on another person full-time. That's not a good solution for us, though, because we have developed a very particular set of procedures and use an unusual mix of software to run our machines. And there's also huge value to having instant communication between the shop floor and the engineer so that problems can be solved as soon as they arise. Managing an outsourced engineer would be a nightmare. Where would I find someone who really understands our operations? And what use would it be if I couldn't control the person's schedule?
I would much prefer to cross-train one of my existing workers. I have several bench hands who already know how we make our tables, who have experience operating my machines and who want to do the job. For the sake of this article, I'm going to morph all of these guys into a single candidate. Let's call him Bill.
After being trained in traditional woodworking, Bill graduated from a technical college in 2000, and he has worked at my shop ever since. He has superb bench skills and seems to be competent at operating our sophisticated machines, but he has had no experience with our computer-aided design and computer-aided manufacturing software. I see him using our computers to browse the Web at lunch, so I know he understands the basics of operating the mouse and a browser, but I don't see him do much with the keyboard and I suspect that his typing speed is slow, maybe 20 words per minute. What I'm saying is, he's great on the shop floor but I really don't know whether he'll be productive working in front of a screen. Still, his understanding of how we work can't be beat, so I'm willing to see how it goes.
I pay Bill $24 an hour (like Bill himself, this is a made up number but a good approximation of the wages I pay experienced bench hands). With taxes and benefits, he costs me $30 an hour. When he's working at the bench, I can sell his labor for $80 an hour. If I move him into the office, no product will be built. So that costs me $80 an hour in lost production, but I'm still paying him his wages. That means the cost of his training is already $110 an hour.
Of course, someone has to train Bill. That would be my engineer. Guys like him cost about $48 an hour, including taxes and benefits. We bill him out at $90 an hour. If the engineer stops doing productive work and starts training Bill, I forgo both his production and the revenue we charge for it -- but I still have to pay him for his services. That's another $138 an hour. So the initial cost of training Bill is $248 an hour.
How long will that training take? Can we teach him everything he needs in a week? Hmmm. It would be nice if we could. If we accomplish that, getting him up to speed would cost me $9,920. In reality, 40 hours of concentrated attention from the engineer will probably get him to the point where he can be a little bit productive, and then he'll need a lot more time to truly master the procedures. My guess, based on my experiences with other workers, is that three months of practice, on real projects, would bring him to acceptable productivity.
At that point, for the sake of our financial model, his cost drops back to his wages ($30 an hour) plus the lost production ($80 an hour) less a percentage of the output of a fully trained engineer (let's say half, or $45 an hour). So while he's getting up to speed, he costs $110 an hour and makes me $45 an hour, resulting in a net cost of $65 an hour.
A three-month training period (13 weeks) of 40-hour weeks will cost me $33,800. Add that to the $9,920 I spent on the first week of training, and you get $43,720. If I find a worker to replace Bill on the shop floor, I have the potential to do more production to cover some of that cost. But in any case, it's a lot of money out of my pocket.
Will Bill demand a raise after I have paid to upgrade his skills? Probably. Engineers make more money than bench workers. And his new knowledge is lodged in his head. I can't take it back. So he has the opportunity to renegotiate our relationship. He can threaten to leave and see what his package of skills is worth on the market. I'll have to bargain with him to get him to stay.
One of the interesting aspects of this scenario is that I have little incentive to bring my shop operations into alignment with standard industry practice. The more idiosyncratic our infrastructure, the harder it is for workers to take their skills elsewhere. One way this plays out is in the mix of software that I use for manufacturing.
Back in 1997, I started to do computerized drafting on a program called PowerCADD. I was using Macs at the time and still do. The standard drafting software in our industry is Autocad, which runs on Windows. Every architect's office uses it, but we don't, which is both a blessing and a curse. I have a substantial library of drawings that we can only open with PowerCADD. But PowerCADD is produced by a very small company, of uncertain stability. At some point it could fail, leaving me stuck with an obsolete library.
On the other hand, I can be more certain that an investment in training my workers in PowerCADD will pay off for me -- knowledge of that program is not worth much on the open market. So I get the benefits of a capable software program without the risk of its operators jumping ship.
To recap: bosses incur real costs when training their workers. And additional training has the potential to upset the balance of power between labor and management. Is it any wonder that some bosses hesitate to make this investment?
Paul Downs founded Paul Downs Cabinetmakers in 1986. It is based outside Philadelphia.