So I expect we have all heard the phrase ‘Economies of Scale’ and have a view on what is meant.
The phrase is probably covered within the first pages of ‘Economics 101’ and every ‘Beginner’s book of management’. I think the idea has even leaked out of these domains and is used in every-day parlance. It is seen merely as ‘common sense’*.
(* please read and reflect upon a hugely important quote on ‘common sense’ when you get to the end of this post)
So what is the thinking behind ‘Economies of Scale’?
Let’s start at the beginning: Why is it said that we benefit from ‘economies’ as an organisation grows larger?
The idea in a nutshell: To run a business you need resources. As you grow, you don’t necessarily need a linear increase in those resources.
Basic example: A 1-man business premises needs a toilet (if he needs to go, well he needs to go). But when the next person joins the growing company he doesn’t get his own personal toilet written into his ‘remuneration package’. No, he has to share the existing toilet with his fellow employee. You can see this logic for lots of different things (one building, one IT system, one HR manager….), but I reckon a toilet is about as basic as it gets.
The theory goes that as the volume of output goes up* then unit costs come down (where unit cost = total cost/ units of output).
(* I’m writing generally now…I’ve moved on from toilet humour 🙂 )
It should be noted that the classical economists that came up with the theory did accept the idea of ‘diseconomies of scale’: that of costs rising as growing organisations become more complex, more bureaucratic…basically harder to manage.
You’ll likely see all this expressed in economics text books with a very simple diagram (below) and, voila, it is surely so!
Getting into more specifics about the phenomenon, three distinct reasons are given for those scale economies:
- Indivisibility: Some resources aren’t divisible – you can’t (easily) have half a toilet, a quarter of a receptionist, 1/8th of a manager and so on.
- Specialisation along with Standardisation: this reason goes way back to the writings of Adam Smith and his famous book called ‘The Wealth of Nations’ (1776). In it, he used the example of a pin factory to explain the concept of ‘the division of labour’. He explained that one person performing all the steps necessary to making a pin could perhaps make only 20 pins a day but if the pin-making process were broken up into a series of limited and standardised operations, with separate people performing them in a joined-up line, productivity could rise to thousands of pins per day per worker.
- Machinery: Investing in ever larger machines mean that they can turn out more and at a faster rate…and our beloved unit costs come down. In service organisations the equivalent could be a ‘bigger, better’ telephone system, IT system,…etc.
Sounds like a water tight case to me – ‘Economies of scale’ proven, case dismissed!
Not so fast…a few dissenting voices:
“All the above seems to be about managing our costs? We are concerned about where this might lead – shouldn’t we be first and foremost focused on delivering value to our customers?”
“We’ve got really low unit costs at lots of our activities…and we keep on making ‘economies of scale’ changes to get them even lower…but this doesn’t seem to be reducing our total costs (they remain annoyingly high)…are we missing something?”
“Gosh, that ‘economies of scale’ average cost curve looks so simple…so all we need to know is when we are at the optimum size (Q) and stop growing. Easy! Can someone tell us when we reach that point? How about a nice warning signal when we are getting close? What do you mean it’s just ‘theoretical’ and no-one actually knows?!”
“I’ve heard that ‘behavioural economics’ is debunking a central assumption within Adam Smith’s classical economic ideas. Apparently we are all human beings (with our own unique purposes), not rational robots!” (Nice link: Who cooked Adam Smith’s dinner?)
“We don’t make pins. We are a service organisation. We have much variety in demand and our customers are ‘co-producers’ within our process…specialisation and standardisation can do much harm to them, and therefore us!”
Meanwhile, on another planet…
Taiichi Ohno developed the Toyota Production System (TPS). In so doing, he used totally different thinking, with profound results.
(Note: Historians have identified a core reason for this difference in thinking as the heavily resource-constrained context that Japan found itself in after the 2nd world war. This was in complete contrast to 1950s America that had an abundance of resources and booming customer demand. In short, Ohno had to think differently to succeed.)
The big difference – Flow, not scale, as the objective: Ohno concentrated on total cost, not unit costs. He realised that, first and foremost, what matters is how smoothly and economically a unit of demand is satisfied, from initial need through to its completion (in the eyes of the customer).
The flow is everything that happens between these points and, as well as all the value-adding steps, this includes:
- all the time that nothing is happening (a huge proportion of a traditional process)
- all the steps that occur but shouldn’t really need to (i.e. they are non-value adding);
- all the repeat and/or additional steps needed because something wasn’t done right; and (the worst of all)
- everything needed to be done when the customer returns with the good or service as not being acceptable (where this could be days, weeks or even months later)
There’s no point in a particular activity being made ‘efficient’ if this is detrimental to the flow.
‘Economies of scale’ thinkers (and their management accountants) are obsessed with how much each activity costs and then targeting reductions. Their belief is that, by reducing the costs of each activity, these aggregated savings will come off the bottom line. Such thinking has led to:
- ‘large machine thinking’ (which also relates to centralisation/ shared services);
- ‘batch thinking’ to make these resources work (allegedly) more efficiently;
- ‘push thinking’ to keep these resources always working – high utilisation rates are king; and
- inflexibility due to highly specified roles and tasks
…which cause a huge amount of waste and failure demand.
Ackoff made incredibly clear in his systems TED talk (using the automobile as his example) that trying to optimise the components of a system will not optimise the system as a whole. In fact, the reverse will be true and we can expect total costs to rise.
Rather than trying to get the cost of a specific activity down, Toyota (and other system thinkers) focus on the end-to-end horizontal flow (what the customer feels). This is a different (systemic) way of thinking and delivers far better outcomes.
It is no coincidence that Ohno is also credited with much of the thinking around waste. It is only by thinking in terms of flow that waste becomes visible, its sources understandable, and therefore its reduction and removal possible.
In short, Cost is in flow, not activity.
Flow thinking has led the design of systems to:
- ‘right-size thinking’ and ‘close to customer thinking’;
- ‘single-piece flow thinking’;
- ‘pull thinking’; and
- handling variety ‘in the line’ thinking (Note to self: a future post to be written)
These all seem counter-intuitive to an ‘economies of scale’ mindset, yet deliver far better outcomes.
(How) does this apply to service?
Okay, so Ohno made cars. You might therefore question whether the above is relevant to service organisations. Here are examples of what the ‘Economies of scale’ mantra has given us in service, broken down into comments on each of specialisation, standardisation, centralisation and automation:
Specialised resources: Splitting roles into front, (middle) and back offices; into demand takers (and ‘failure’ placators), transactional processors, back room expert support teams and senior ‘authorisers’…meaning that:
- we don’t deal with the customer when/ where they want;
- causing delay, creating frustration – which needs handling;
- incorrect setting of customer expectations;
- unclear ownership, leading to the customer having to look out for themselves
- we have multiple hand-offs;
- causing batching, transportation, misunderstandings, re-work (re-reading, re-entering, repeating, revising);
- we break a unit of value demand into separate ‘work objects’ which we (hope to) assign out, track separately, synchronise and bring back together again (…requiring technology);
- we collect information to ‘pass on’ (…requiring technology)
- often passing on incomplete and/or incorrect information (or in Seddon’s words “dirty data”), which escalates to the waste of dealing with the defects as the unit progresses down the wrong path;
- we categorise, prioritise, allocate and schedule work around all these roles (…requiring technology)
- …all of the above lengthens the time to deliver a service and compromises the quality of the outcome, thus generating much failure demand (which we then have to deal with)
Standardised activities: Trying to achieve a standard time (such as Average Handling Times) to perform a standard task (using standard templates/ scripts) that appears to best fit with the category that ‘we’ (the organisation) jammed the customer into
- rather than listening to the customer’s need and attempting to deliver against it (i.e. understanding and absorbing customer variety);
Centralisation: Seeing ‘shared services’ as the answer using the “there must be one good way to do everything” mantra.
- creating competition for shared service resource between business units and the need for SLAs and performance reporting;
- requiring some ‘super’ IT application that can do it all (“well, that’s what the software vendor said!”);
- ‘dumbing down’ the differences between services (and thus losing the so-called ‘value proposition’)
- loosening the link between the customer and the (now distant) service.
Automation: Continually throwing Technology at ‘the problem’ (usually trying to standardise with an ‘out of the box’ configuration because that will be so much more efficient won’t it) and, in so doing, creating an ever-increasing and costly IT footprint.
Whilst technology is amazing (and can be very useful), computers are brilliant at performing algorithms (e.g. calculations and repetition) but they are rubbish at absorbing variety, and our attempts at making them do so will continually create failure demand and waste.
In summary: ‘Economies of scale’ thinking is more damaging in service because of the greater variety in demand and the nature of the required outcomes.
This post isn’t saying that scale is wrong. It is arguing that this isn’t the objective. Much harm is, and has been, done by blindly following an activity focused logic (and the resultant ‘specialise, standardise, centralise, automate’ mantra)
Further, I get that some of you might say “you’ve misunderstood Steve…we aren’t all running around saying we must be big(ger)!”…but I’d counter that the ‘economies of scale’ conventional wisdom is implied in a relentless activity cost focus.
Put simply, “Economy comes from flow, NOT scale” (Seddon)
Beware ‘Common sense’:
“There is a time to admire the grace and persuasive power of an influential idea, and there is a time to fear its hold over us.
The time to worry is when the idea is so widely shared that we no longer even notice it, when it is so deeply rooted that it feels to us like plain common sense.
At the point when objections are not answered anymore because they are no longer even raised, we are not in control: we do not have the idea; it has us.” (Alfie Kohn)
Credit: The ‘Economies of scale’ explanation comes from reading a John Seddon paper.
Being fair to Adam Smith: He understood that the specialisation of tasks can lead to “the almost entire corruption and degeneracy of the great body of the people [the workers]. … unless government takes some pains to prevent it.” i.e. it might be great for the factory owners…but their workers are people, not machines.