Variety and variation — their role in business
Recently, I overheard a conversation, in fact it was more one way than a conversation would normally be. The person speaking was instructing her listeners to take a blank piece of paper, close their eyes and follow her instructions as she proceeded to tell them to turn, tear, and fold their sheets of paper various times and in various ways. What struck me, was the differences between the end result created by the many people following these simple instructions. What I saw was ‘variety’, at least, that was my interpretation. But, here’s the rub… the instructor called them ‘variations’. This piqued my curiosity. What is the difference between ‘variety’ and ‘variation’?
A simple glance at the English Oxford dictionary provides the following definitions:
variety (noun) — the quality or state of being different or diverse, the absence of uniformity or monotony.
variation (noun) — 1. a change or slight difference in condition, amount, or level, typically within certain limits; 2. a different or distinct form or version of something.
My perception, that this was a demonstration of variety, fits the first definition well. There was no uniformity among the folded and torn sheets of paper and they all were in a different state in terms of how they looked. However, if we consider the outputs from the point of view as ‘variation’, we can also argue that we have different and distinct versions of a piece of paper, each with its own unique set of holes and creases. Therefore, I would argue that we were both correct — they showed variety and variation.
But I cannot end this argument here. Variety and variation are different, and I think it’s important to make a distinction between them to understand how we interpret variety and variation around us, particularly in how we understand quality.
The first time I heard anyone mention that variety and variation are distinct in the context of business quality control was while listening to John Willis’ excellent Profound podcast. This podcast series is a tribute to the statistician Dr. W. Edwards Deming. The culmination of Deming’s lifework is his System of Profound Knowledge in which he describes four pillars of business: appreciation of a system, theory of knowledge, psychology and…. knowledge of variation. There’s that word again!
Deming states there are two types of variation: common cause variation and special cause variation. By understanding the difference between them, we can improve quality and reduce unnecessary waste by recognising changes that fall within boundaries of the system and those that fall outside those boundaries. If you want to learn more about variation in this context, I recommend reading Deming’s book The New Economics for Government, Industry, Education (MIT Press, 3rd edition, 2018). What I’d like to emphasise here though is that variation is measured over time — the state of a system changes from one time period to another. In other words, variation is temporal. If we measure the state of a system, for example, the number of vulnerabilities identified within source code following an application security scan, that number will change between scans. It may go up, it may go down. It may even stay the same. This is variation. Although, we may also notice a variety of different types of vulnerabilities from the same scan. And without trying to confuse matters, if the variety changes, then it too is demonstrating variation.
When we consider variety, I think about the choices we have. Henry Ford famously declared that you can have any colour of his Model T Ford car as long as it was black. There wasn’t a choice at all. There was no variety. In today’s motoring world, we have plenty of choice. We can choose multiple styles of cars, in many different colours, with may different configurations. This is what we mean by variety.
Variety serves a purpose — we create a variety of products to serve our customers’ needs and gain a competitive edge. Indeed, our customers also have variety — they have different requirements based on their preferences, budget, and lifestyle. According to Ashby’s Law, only variety can absorb variety. This is known as Ashby’s Law of Requisite Variety. It means that we must either develop products to meet the variety of our customer (amplification), or reduce demand of the customer to meet our product variety (attenuation). In Ford’s case, he reduced customer variety by reducing customer expectations to only needing a cheap, mass-produced (and black) car. Therefore, he was able to mass-produce the same style (and colour) of car to meet that expectation. Nowadays of course, car manufacturers produce a variety of quality cars cheaply thanks to more efficient ways of working. This variety was met by Shingo Shigeo’s Single-Minute Exchange of Die (SMED) technique that allows manufacturers to produce a wide range of products swapping machines that make those products very quickly.
Measuring variation and variety involve different sources of data. Variety is simple to measure, it is a count of the number of items that meet the criteria of being different or diverse. Variation is not so easy to measure. As mentioned, we have common cause variation and special cause variation which are governed by upper and lower limits of variation. These limits are calculated based on three degrees of deviation from the mean. Deming demonstrated this very well with his red bead experiment. We need to measure variation so we can see when the system itself needs changing. Too often, we react to variation without fully understanding what caused that variation. For example, if we see an increase in the number of vulnerabilities identified during a scan, managers act to find out why there’s an increase. If the number falls following the next scan, managers see this as a success. Yet, in most cases this variation falls inside the boundaries of common cause variation, the manager’s reactions are tampering and have no discernible effect on the system as a whole. As a result, we need to be careful how we measure variation if we want to understand how our systems are performing.
In summing up, variety and variation are important measurements and we need to understand them. When I think back to that paper tearing exercise, I see variety and variation. What is important is that we make a clear distinction between them both and understand their implications.
