Thoughts on Cost Accounting

I'm just going to dive right in. Cost Accounting is hard. It really sucks, and there is no easy way to shortcut the problem. In the most basic form, we quickly hit an information limit. You can't know something unless you track it. Tracking it costs you something, but how much would it cost you if you didn't track it? Who knows because you didn't track it!

Indirect Costs Suck

If a company forecasts sales and equipment utilization incorrectly, they will be selling an hour of their time now, for less than that hour actually cost them in the context of the entire year's sales. To combat that, you have to review actual vs forecasted utilization often, but not too often for fear of false readings.

Fixed costs are difficult due to the need to forecast and budget. Say a machine expects to have a 10,000 hour lifetime. You forecast you will be able to bill 1000 hours a year on this machine. If it costs 100k, then we should expect a cost of $10/hr. This would give us an hourly cost that would pay back the machine over 10 years. Simple, right? What happens when the loan is only for 5 years? Just charge $20/hr right? This is 100% markup over your 'cost'. Add to this, that you can now depreciate the entire cost of this machine in the first year for tax purposes. What is your 'cost'? What happens when the machines life only makes it to 8000 hours? Should you change your pricing year to year based on what you need for cash flow? Customers generally want predictable prices so they can plan more easily.

So what is the 'right' solution? There really isn't any single answer. It really comes down to what works best for your situation, in your market, with your product, at that point in time. The best way to approach this is track everything (utilization) and sell your product at a competitive market rate. Generally, cash flow is the limiting factor on how low you can set your rate. You just have to be aware of the many factors involved before you claim to know your profit margin. Generally, you can only truly know your profit margin after the fact. When the machine 'dies'. If it lasts 8000 hours instead of 10000, then, let's say you figured degradation was higher than what it should be by hour 5000. That means you already billed 5000 hours at only 80% of what you should have been. So do your charge a 66% premium on those last 3000 hours? You can see how after the fact is really the only way we can get a good approximation of 'cost'.

Industries with high fixed costs need to carry a larger margin for these unexpected expenses. Generally, they always sell their assets before they wear out (usually after tax depreciation has been consumed). They are able to 'pick' when their machine 'dies'. They are just shifting the unforeseeable breakdown costs to whoever buys the used machine. These markets end up in sort of two tiers, people who buy the new machines, and people who make their margin in buying and dealing with the unknowable future breakdown costs of used machines. Any product that has steady demand and a decent margin are usually the product paying for the new equipment. Any non-steady demand products generally end up being produced on used equipment. The margin is variable due to the unknowable, and unforecastable breakdown and repair costs. Basically, the more you pay the more certain your costs. This is a generalization, and by no means a hard and fast rule.

Nested Cost Structures

Another fun perk with cost accounting is that costs you are determining are set inside a larger cost structure (ie., Process > Product > Company > Country > World). In the above example, if the machine was subsidized by a government loan what is the cost? How do you factor in all the taxes that were paid by individuals and companies? If the government is running a deficit, then the taxes aren't even paying for the subsidization, or the government itself. Does that mean taking a subsidization from that government technically increases your 'true' systemic cost, while the market (people buying the product) actually see a lower price than if there were no subsidy? The more cost accounting one does, the less there is ever truly a 'cost', but rather a series of statements with a time-frame of validity which create an illusion of 'knowing' the cost. Usually, just long enough to 'run the numbers' through our spreadsheets and get a product to market.

A waste of time?

The difficulty when you try to 'get to the bottom of it', is that you end up actually doing a decent job of forecasting and weighing all the variables that will allow you to budget a sustainable business. Depending on your market, you can often find you just priced yourself right out of it. So what was the point of figuring all that out? Well, you basically tried to predict the future of your company, priced what it would take to get it there, and you added more items to your cost structure to increase accuracy and build in scalability. That doesn't sound like a waste of time at all. You now have a few thoughts to ponder.

  • Is my predicted path the financial path I should take my company?
  • Do I know if my predicted path is even viable (untested assumptions)?
  • Are there other ways to accomplish my goals with less cost?
  • What part of my 'new' cost structure shows me something I hadn't previously thought about?

Fast forwarding time in your brain is quite a useful exercise to do. Why wait for the problem to arise in the real world? If you run enough scenarios in your head, you might be able to know when to bounce between them to maximize your current situation's potential.


Doing too much cost accounting is wasteful, and doing too little is dangerous. The key is lowering the cost of tracking information. The more data you can gather the better. Maybe one day we can cost account the full stack, from energy all the way to the nation-state. However, I don't think that is even possible. I honestly don't think there is enough data collected to get an accurate number.

I hope after reading this you feel less alone when you are overwhelmed while looking at your business finances and trying to understand where things slot into your expense sheet. Now go run some numbers!

Julian Stahl

I like to be accurate. I try very hard to understand things from many perspectives and make sure the ground I stand on is solid.

Next to some mountains in Colorado