Let me declare today that I have a new enemy - Cost Accounting.
In 1981, at a conference of Management Accountants, Eliyahu Goldratt declared Cost Accounting to be "Enemy #1 of Productivity". In my work as a consultant for ThoughtWorks, I value productivity highly. Hence the enemy of my friend is also my enemy.
Goldratt, author of the popular business novel "The Goal", showed how traditional cost accounting easily and trivially leads intelligent managers and other decision makers to make bad decisions - decisions that have massive negative effects on performance and profitability. This was actually not surprising news, as the problems with cost accounting had been recognised. But Goldratt also proposed an alternative.
His alternative was titled "Throughput Accounting", and was described in the 1998 book by Thomas Corbett. It is an accounting model which does not focus primarily on local cost optimisation and instead focuses primarily on overall system throughput. Importantly, it does not attempt to allocate overhead costs across products and services.
Why is this important? Cost accounting attempts to allocate a companies fixed costs over a given time period to the products or services produced in that period. It enables managers to ignore the fixed costs and focus instead on the results of each period in relation to the products produced. This made sense in the 1890s, when cost accounting was developed, as most companies had extremely low fixed costs. The majority of costs were labour, which at the time was fully variable, as workers were only paid for the hours they laboured at the demand of the employer. Although there was distortion due to the overhead allocation, it was extremely small.
Modern businesses, however, have a much lower proportion of fully variable costs. Most costs are fixed, especially labour as now most employees work fixed hours regardless of demand. Due to this environmental change, the distortion caused by cost allocation is much larger and can cause management decisions, based on cost accounting, that directly result in large loses of productivity and profit.
As an example, imagine a company has fixed overheads of $10,000 (plant and labour) per month and currently produces 800 widgets with a materials cost of $40. The unit cost per widget would be $52.50 ($40 + $10,000/800). If an order now came for an additional 500 widgets at a price of $50, then management based on cost accounting should reject the order as it would result in a loss of $2.50 per widget. However, as the majority of costs are fixed, the increase in production would actually result in a new unit cost of $47.69 ($40 + $10,000/1300), so profits would actually increase by $1,153.85. This might seem clear, but imagine this example with many different products and prices in the mix and quickly it becomes hard to see the mistake. This is only one example of a poor decision made due to a cost accounting metric.
So why am I so concerned? In the software industry, especially, there is very little variable costs. The overwhelming expense is for developers, which are typically in short supply, so they are not a variable cost. Basing our decisions around software projects and the ROI from software delivery on cost accounting can lead to incredibly poor outcomes.
In future blog posts, I'll also explore what I believe is another critically important aspect of Throughput Accounting - the focus on throughput and inventory reduction over costs. I believe this has the potential for a massive effect on the software delivery process, especially when combined with the Lean practices that ThoughtWorks is pursuing in the industry.
Believe me when I say this - Throughput Accounting will become of major importance in our industry. It will make a dramatic change to the way we finance software projects and it has the power to put real numbers around why we use Agile and Lean principles and practices (as shown in David Andersons book "Agile Management for Software Engineering").
If your interested in another good introductory text, in addition to Corbetts book, try the book by Steven Bratt, also titled "Throughput Accounting".
Friday, July 18, 2008
Subscribe to:
Post Comments (Atom)
3 comments:
Fascinating post, I definitely want to learn more about throughput accounting.
Although from what I dimly remember from financial accounting class in college, widely allocating fixed overhead costs is generally considered a bad idea even when using traditional accounting.
My favorite example is the fixed allocation cost death spiral, where the "unprofitable" stores/products are killed, causing the costs to be distributed amongst n-1 stores/products. The relatively increased allocated fixed costs will then make other stores/products appear unprofitable, so they too are closed, and so on...
and so on...
I am not convinced. Cost accounting was specifically developed to to deal with fixed and variable cost, because this was what industrialisation meant - huge upfront investments in plants and machinery (the historical argument in the wikipedia article seems strange to me).
There is a specific concept in cost accounting that helps you decide whether to sell or not, that should be used in your example. It is called contribution margin and defined as Sales Price - Variable Cost. As long as this is greater zero it helps covering your fixed cost. In your example it would be $50 - $40 = $10. The trouble is that this price is only good enough, when you sell enough to cover all the fixed cost. The actual crux in practive is to determine the actual variable and fixed cost, because in the long term all costs are variable and the variable cost is usually not linear, so your model works only around a certain operating point.
I am curious to learn what throughput accounting is really about.
Cost accounting is a very powerful tool and as any tool to be used responsibly.
Apparently most financial/management accounting classes do tell you that making decisions based on full-burdened costing is a bad idea, and yet this practice permutates all sorts of corporate decision making. It's very hard to escape when your measures of success are built on cost-accounting metrics.
Contribution margin is also a great concept for solving the sort of problem I used in my very simple example. In fact, it's almost the same as the concept of 'Throughput' from Throughput Accounting. However it certainly doesn't solve all the problems of Cost Accounting - especially when it gets used in conjunction with all the other CA measurements and metrics. Conversely, TA has other measures and metrics that do use it much more wisely. I'll certainly explore Throughput (aka contribution margin) in more detail in my next post on the topic.
Post a Comment