I’ve been up to my eyebrows in business simulation preparations for my last MBA subject (woohoo!) and it got me thinking about ROA and current trends in IT. The simulations have been about manufacturing companies (cars, now sensors) where ROA is an important measure of company performance. I’d argue it always is.
ROA stands for Return on Assets. The Wikipedia article on ROA is pretty good, and this article at Investopedia is not bad either. It’s a measure of how much money (Net Income) you make from a certain amount of assets (Total Assets). Certain industries are capital intensive, like mining and manufacturing, because you need a lot of assets (machinery, factories, etc.) to make whatever you’re selling. Others, like consulting, or Internet start-ups, aren’t capital intensive because you don’t need much physical gear. It’s used as part of the Du Pont model for financial analysis, which I like quite a lot.
The thing about ROA is that it goes hand in hand with your overall business strategy. There are only two ways to get ROA to go up: make more money with the same stuff, or use less stuff to make the same amount of money. And there only a couple of ways to make more money: sell more stuff, or sell the same stuff at higher margins. And, once again, higher margins can be done only in two ways: lower your costs, or raise your prices. And if you lower your fixed costs, you improve ROA directly by making the same amount of money with less stuff.
In an ideal world, you’d do all of them, but in the real world you’re constrained as to what you can do by a bunch of things: which industry you’re in (manufacturing vs. consulting), your position in the industry, overall economic trends, etc. That’s why lowering costs is so appealing to average managers: it’s obvious and almost always available as an option, so you don’t have to do any thinking before you use it.
History of Manufacturing
IT is changing. It looks a lot like the way manufacturing changed from handicrafts to automated moving assembly lines. Adam Smith wrote about the division of labour in his 1776 classic The Wealth of Nations, where he wrote at length about the manufacture of pins. Traditionally, one person would make the whole pin, but he wrote about how by dividing the job between multiple people (one person to make the body of the pin, and then have a second person make the head of the pin) the two people working together could make more pins in an hour than two people making the whole pin themselves.
The process of one person making a whole item was called the English system of manufacturing. It was based on skilled machinists building all the parts for a given item (turning), and fitting them together. Hence the term “fitter and turner”. Things were done this way because, despite the skill of the machinists, parts made by different people weren’t identical, so they wouldn’t fit together properly with parts made by someone else. It’s not a big deal with pins, but it is with muskets and rifles, which tend to explode and kill your soldiers if they’re not built properly.
The introduction of standardised, interchangeable parts that could be assembled together was a massive advance, which actually happened in England the early 1800s, but it didn’t really catch on until the idea was re-imported from America in the mid-1800s. It’s now known as the American System of Manufacturing for that reason. It meant that different people would build parts, and then check them against specifications (with tolerances). If they passed, you knew your parts would fit with parts built by other people. And now you can have not just different people building parts, but different companies. Possibly in different countries. It’s division of labour writ large.
I’d argue that IT is about midway through the transition from the English system to the American system of manufacturing. There’s a lot of hard-crafting going on, and lots of almost-but-not-quite identical builds. Some more advanced places are doing better, and this is where ROA comes in.
If your company has 10 people, hand-building their desktop machines is easy enough. They’re all built slightly differently, because you learned a new technique for the second five, and made a mistake on a couple of them, and they all have slightly different needs. Fixing them if they break requires a tailored approach because of these differences. But it’s not a big deal, because there’s only 4 and they don’t break that often. Let’s say you have 1 full time IT person to support this setup.
But now imagine your company has 10,000 end-users who all need desktops or laptops. If you use the same techniques as for your 10 person company, you’ll need 1000 IT people to support your end users, and that ignores any other management or HR overheads. That’s expensive. This is exactly the same problem as manufacturing in the 1800s. Ford wouldn’t have been able to build 941,042 cars in 1920 if not for the moving assembly line. And in ten years, the price of the Model-T had dropped from $850 to $260. That’s a 70% price drop.
The pressures on IT are very similar: businesses want more, and they want it cheaper. You don’t have the “raise prices” lever available, so you have to cut costs. How do you do that? Well, you can make less, but the business wants more, so that lever isn’t available. You have to get more efficient. How do you do that? By churning out more things with less effort. The only way to do that is with standardised, interchangeable parts. And if you can automate the process, so much the better.
That’s why the more advanced places are automating everything. They are the Japanese to your mid-80s American car company. It’s also why their stuff is better and cheaper than yours, and why you have to change. If you don’t, you’ll be out of business.
Human Cost of Change
But what does this mean for the people in IT? Just like in manufacturing, there will be less work available for skilled machinists; semi-skilled labour will be just fine. And as more things are automated, there’ll be even less work for semi-skilled labour. Just like modern car factories, where most of the work is now done by robots. Over time, expect wages to drop as demand for these kinds of people drops. We’re already seeing this happen in Australia.
There will still be some IT machinist jobs, don’t get me wrong. There just won’t be as many, and they won’t be doing the work they do now.
I’d be learning how to build or program the robots, if I were them.