Over time, it is easy for processes to corrupt or malfunction and carry on without notice. Malfunctions happen because the nature of work changes. New customers come with different expectations. When and how they expect your services to be delivered changes. A seemingly small change in opening hours or how people communicate, by phone, email or text, ripples out. Subtle differences that don’t seem like much at the time. They might be frustrating if they are noticed at all, but people find workarounds, because that’s what people do. The work must be completed despite the odd change. But then processes drift, inefficiencies creep in. No one has done anything other than try to get through the work—things just don’t run as smoothly as they once did.
These inefficiencies are inevitable, but were once much easier to absorb than they are now. Junior employees who carry a lot of the routine process work were relatively inexpensive compared to more senior hires. But within the last five years, the cost of employing one full-time worker at the statutory minimum has risen by roughly 50% when you include employer’s national insurance and pension contributions. In April 2021 the cost of employing someone on the minimum wage for a 40-hour week was £20,239. By April 2026 that had increased to £30,258. For an operation with 50 junior roles that is just over half a million pounds of additional expense each year – a profit-shattering increase in junior salaries alone.
But the problems for business owners do not end with junior salaries, because remuneration for people in senior positions must be increased too. If not, what is the point of their knowledge, experience and judgement? And that wage compression will not just be felt by the more senior workers; the people in junior positions will wonder why they would ever want to work harder for just a little more pay—let alone take on the disproportionate responsibility.
In practice, the difference between a junior wage and a manager wage narrows while the business carries significantly more cost. The impact on margin is immediate, which then compounds as productivity falls.
The obvious solution is for business owners to increase prices, but that is seldom an option—especially since operations with a broad base of junior roles are generally structured around higher volume, low-margin work. Figures published by the Office for National Statistics show this predicament clearly. Service companies may be charging 21% more in 2026 than they were five years ago—but the cost of employing someone at the statutory minimum has risen more than double that.
Where you must increase wages and cannot raise your prices—the only thing left to move is the operation itself.
The temptation may be for business owners to offshore some of their human operating system where wages are ostensibly cheaper, or to implement AI in the hope of reducing processing costs. Both can deliver real savings when they are implemented well. However, there is a real danger that operational efficiencies just become institutionalised and harder to see. The cost may reduce and be once again less obvious, but it is still lurking out of sight until increasing costs bring it into sharper focus once more.
The nature of the work determines the shape that your human operating system should be—the people you want and the capabilities they need to make your customers happy. The starting point, then, is to establish what work flows through your system and whether its complexity matches the human capability you have built around it. Why the structure of your service operation determines your margin and morale contains a simple question that you can use to find out.
The second consideration is how much of the remaining work inside the system needs to be done at all.
John Seddon pioneered a systems thinking approach to operations design that asked how much demand arises because of avoidable errors—a source of operational strain he termed Failure Demand. Seddon’s decades of experience suggest that failure demand can run between a third and two thirds of all demand experienced by service operations, which is a significant opportunity for cost-saving. Failure demand may manifest as clients calling to chase progress or in work that needs to be redone because when it was first done it was incomplete or defective. Errors can crop up because the complexity of the work is not matched to the capability of the human operating system, but also through process design.
Some of the problem, which Seddon also argues, arrives with traditional views on how the work should be managed. Most solicitors or insurance companies will be familiar with managing the demand from customers who call in to report their accidents. Service agents in turn record those incidents as new claims. Attempting to ensure flow, service agents are often targeted on completing a claim within a given time, which in turn gives supervisors some comfort that they might be able to predict demand by handler. But claim reports (like all customer enquiries) vary. Some are simple and can be completed quickly. Others are complex and take some time. It is not just complexity—some customers prefer full explanations whereas others are happy with succinct guidance. But by designating time as the objective, the service agent becomes the problem if the target is missed. No allowance is made for variability.
The service agent then has a couple of options. They can abandon hope of ever hitting target on those claims that take longer. Or they can rush through parts of the claim report and pass the problem down the line. Invariably targets are met and the substandard work causes a problem for someone else in another department. Besides creating more stress for the operation, the results create delay and involve speaking to the customer again to ask questions that have already been asked. This is how interdepartmental silos begin to form. Where customer dissatisfaction comes from. And why people turn up and struggle with morale. All because of an arbitrary target that cannot reflect the variability in the nature of the work and doesn’t help it move through the system.
The solution is to regard the operation as a complete system regardless of whether it is sectioned into departments. Decide what the operation needs to accomplish to be successful and create a meaningful measure of performance. The life cycle of a claim would be appropriate where speedy resolutions are expected by customers, a measure that transcends departmental boundaries. Such measures can be plotted to account for variation within reasonable parameters. Although demand varies, most of it will cluster. Where it doesn’t—that’s where you look. Redesigning the system stops the same issues cropping up again. Operating costs reduce and morale improves as people no longer have to work around problems.
These inefficiencies rarely manifest in complaints, because people cover up problems and customers can be forgiving. Accordingly, leaders might obtain some insight from complaints, but never a full picture. In fact, if each department along the value chain has its own target, and they all do whatever they need to do to achieve it—the operation may appear to perform perfectly well. Nevertheless, it carries the cost of inefficiencies.
There is a quick way to assess failure demand—listen to calls and check the work in progress. Any task that involves chasing suppliers or rework is a cost the business could be saving. Every call from a customer wanting to know how their service is progressing is too—it is also a signpost to a broken process further upstream.
What the Operations Triangle and a systems approach share is a need to understand the nature of demand and how the system should be designed to meet it. Unlike traditional targets which measure performance in units of time or money, neither considers complexity or variation within the system as bad. In fact, they enable you to design your operation to work despite its variations while reducing waste and any associated costs.
The solution to rising wage costs is to eliminate work created by the system because something has gone wrong. Not by restricting demand, which would reduce turnover. Not by reducing headcount, which would just create even more problems meeting demand.
You just have to stop doing work that no one is paying you to do.