How to improve business operations without a major restructure.

When an operation becomes difficult to manage, a restructure can feel reassuringly decisive. New teams, new reporting lines, a diagram that finally fits on one slide.

Then Monday arrives and the same approvals still wait in the same inbox.

Sometimes the structure genuinely needs to change. The question is whether it is causing the problem you need to solve. For growing businesses, proportionate operational improvement starts by understanding what works, where performance breaks down and what intervention the evidence supports.

1. Why changing the organisation chart may not change the work

Moving the kitchen table does not fix an oven that only works when someone holds the door shut.

It may improve the room. Dinner will remain unpredictable.

An organisation chart describes reporting relationships. It does not fully explain how work enters the business, which information is available or who can make a decision.

A restructure may help where authority, expertise or coordination is fundamentally misplaced. But it can leave process defects untouched, while introducing transition costs and new dependencies.

I’d test the proposed mechanism: how will this reporting change remove the cause of the problem? If the answer is simply “people will work better together”, the intervention needs a more concrete explanation.

2. Identify what already works

A business with some operational friction is not necessarily an operation that needs replacing.

Even a temperamental car may have perfectly good tyres. Buying four new ones because the radio fails would be an interesting maintenance strategy.

Look for work that completes reliably and understand the conditions supporting it. The right inputs, accessible expertise, clear decisions or a dependable routine may already exist in part of the organisation.

Compare those conditions with the areas under pressure. Differences can suggest practical improvements without requiring wholesale redesign.

Be careful about exceptional individuals who conceal process weaknesses through extra effort. A result can look reliable while depending on a person continually rescuing it. Ask what would happen during absence or increased demand.

3. Distinguish structural problems from process problems

If three teams cannot agree who decides, there may be an authority issue. If everybody agrees who decides but the information arrives incomplete, a different repair is needed.

Both can create the same visible delay.

Separate the problem’s symptoms from its operating conditions. Examine decision rights, workload, skills, system access, process requirements and measures.

This prevents one category of intervention from becoming the answer to every difficulty. Training cannot supply missing authority. A new manager cannot automatically repair poor data. An extra approval can make a queue worse.

DeaconOps would follow actual work and test which conditions most strongly explain the result. Our guide to finding the real cause of poor performance covers that investigation.

4. Find the smallest change that can be tested

A request repeatedly arrives without a reference number. You could commission a transformation programme. You could also begin by fixing the request form.

The second option deserves a fair hearing.

A proportionate intervention addresses a specific cause with a clear expectation of how the outcome should improve. Examples include changing intake requirements, removing duplicate entry or clarifying a recurring decision.

Small does not mean inconsequential. A modest change at the constraint can affect the whole operation. Conversely, a large improvement in a low-impact activity may achieve little overall.

Specify the problem, proposed change, expected effect and measures before testing it. This makes it easier to distinguish a useful intervention from a generally popular idea.

5. Learn through a controlled trial

Replacing the whole menu before tasting the sauce is a brave approach to catering.

Businesses sometimes apply the same confidence to operational change.

The Plan-Do-Study-Act approach, associated with Deming and developed in improvement practice, offers a disciplined learning cycle. Plan a change and prediction, test it, study the evidence and adapt the next step.

A pilot needs representative conditions. Success with the easiest cases and the most experienced volunteers may not demonstrate readiness for wider use.

Agree how customers and essential controls will be protected, and how to reverse the change if it creates unacceptable consequences. Examine extra work elsewhere as well as the intended saving.

The purpose is learning early enough to adjust, rather than defending a decision because implementation has already become expensive.

6. Involve the people who must make the change work

A process designed without its users can be wonderfully tidy.

Unfortunately, its users may continue doing the work in another spreadsheet.

People close to delivery understand exceptions, timing pressures and dependencies that a process map may omit. Their participation helps test whether the proposed method is executable.

That does not mean every suggestion requires unanimous approval. Leadership still needs to resolve trade-offs and make decisions. But unexplained resistance may reveal a real problem with the design.

Ask what becomes harder under the new method, which task remains necessary and where capacity is supposed to come from. Incorporate the answers into the operating design, training and implementation plan.

7. Recognise when a larger redesign is justified

There are occasions when patching the process becomes the operational equivalent of holding the oven door forever.

A bigger intervention may be the more proportionate answer.

Structural change can be justified where the operating model cannot support the service, responsibilities conflict persistently, critical expertise is fragmented or the business has fundamentally changed.

The case should explain the problem, alternatives, dependencies, costs and transition risks. Include how current service will continue while responsibilities and systems change.

Compare the cost of redesign with the recurring burden of maintaining the present arrangement. Proportionality is about the relationship between intervention and need, not a commitment to keeping every project small.

8. Turn an improvement into normal operation

A successful test is encouraging. It does not automatically update instructions, remove obsolete trackers or tell the night shift what has changed.

Those details are how an improvement survives its launch email.

Assign ownership, update the working method and agree how performance will be checked after implementation. Retire the old route deliberately where appropriate, rather than allowing two competing processes to continue indefinitely.

DeaconOps’s Improvement Projects connect design with delivery. Where the cause remains uncertain, an Operational Diagnostic can establish the right starting point.

Start a conversation if you want to improve the operation but are unsure whether the problem justifies a wider restructure.

Questions about proportionate operational improvement

Can a small business use formal improvement methods?

Yes. The discipline can be simple: define the problem, test a change and review the evidence. The paperwork should fit the work.

Does avoiding a restructure mean leaving responsibilities unchanged?

No. Decision rights and task ownership may need to change even when reporting lines remain the same.

How do we know whether a pilot worked?

Agree the expected outcome and measures beforehand, compare appropriate periods and examine quality, customer impact and work created elsewhere.

---

About the author

Dean Cain is the founder of DeaconOps, with more than 15 years’ experience across operational delivery, leadership and regulated environments. Read about the DeaconOps approach.