Operating expenditure is the recurring cost of running a business, covering people, software, suppliers and overhead. Reducing it means removing or reshaping recurring commitments, which is slower and more political than reducing one-off spend.
What counts as opex, and what does not
Operating expenditure is the recurring cost of running the business: salaries, software subscriptions, rent, professional fees, marketing, logistics. Capital expenditure buys an asset that is used over years and depreciated. The line between them is set by the recognition rules in the AASB accounting standards rather than by preference, and the classification matters because it decides which line of the P&L a decision lands on.
The practical difference is that opex is a set of standing commitments. Cutting it means cancelling something someone is currently relying on, which is why it moves slowly and why it is rarely a purely financial exercise.
The four places operating cost actually sits
Almost every opex programme resolves into four categories. Sizing them before touching any of them is the step most often skipped.
| What usually gets cut first | What the cut actually costs you | |
|---|---|---|
| People | A hiring freeze, then a restructure. | The largest line and the slowest to reverse. A freeze applied evenly also freezes the roles that were about to pay for themselves. |
| Software and cloud | A licence audit and a round of cancellations. | Usually real money and genuinely low risk, but it is a one-off unless someone owns renewals as a standing job. |
| Suppliers and outsourced services | Renegotiate on price at renewal. | Price is the smallest lever. Scope, volume commitments and payment terms move more, and moving terms changes your working capital too. |
| Facilities and overhead | Sublet space, trim travel and entertainment. | Visible, symbolic and usually smaller than expected. Worth doing, not worth mistaking for the programme. |
A sequence that works better than a framework
Lean and Six Sigma both work, and both assume a stable process to improve. Most scale-ups do not have one, so the framework gets applied to a process nobody has written down and produces a diagram rather than a saving. The order that holds up is more boring than that: document how the process truly works, use AI to identify the risks and the timesinks, then automate one small step at a time.[1]
The limit is worth stating up front, because it sets the size of the prize. AI can handle 80 to 90% of basic finance tasks like checking, reconciliation, analysis, forecasting or modelling, but the last 10% requires a qualified human who understands what the challenge is and what good looks like for the outcome.[2] An opex case built on removing that last 10% is a case that fails in month four.
Why benchmark case studies are the weakest input
The search results for cost reduction are full of case studies with a headline percentage attached. They are close to useless for a decision, because you cannot see the starting point, the accounting treatment, what was reclassified rather than removed, or what broke eighteen months later. A saving reported in a different business at a different stage is not evidence about yours.
What is worth having instead is your own baseline: cost per category, per head and against revenue, tracked monthly, so a movement is visible without anyone building a special analysis. That is a reporting problem before it is a cost problem.
None of this reads properly without a P&L someone can actually explain underneath it.
Who owns cost reduction in a growing company
Headcount is the largest opex line in almost every services or software business, so most cost programmes become people programmes whether or not anyone says so. Over the last couple of years significant redundancies, headcount cuts and restructures have created a market where many highly skilled finance professionals are looking, while the volume of available roles has not moved in step.[3]Where reductions do involve roles, entitlements including redundancy pay sit under the National Employment Standards, and they are a cost of the decision rather than an afterthought to it.
There is a specific trap here for founders. Cost pressure regularly stops finance hires from happening even when the business case for them is clear.[4] Declining to add the person whose job is to find the savings, in order to save money, is a decision that looks disciplined on a spreadsheet and costs more than it saves.
If that is the argument you are having internally, here is how to build the business case for a finance hire.
What this says about who you hire
The profile that does this work is not the cheapest one. In the Australian market a half-decent finance hire who can add value beyond compliance might cost $150-160k, while someone with strong experience in automation, tech, AI, capital raises and acquisitions is closer to $200k+, and a top candidate I know is currently on $230k plus super.[5] The gap between those two numbers is roughly the size of the savings the second person is expected to find.
You are not looking for an AI specialist. Employers value candidates who are adaptable and translatable with tech, particularly in automation and process improvement, with a genuine interest in AI, without expecting expert-level capability.[6] Someone who can document a process, see where the time goes and change it one step at a time will take more cost out than someone who arrives with a methodology.
Common questions
What is opex reduction?
Opex reduction is lowering the recurring cost of running a business, as opposed to capital expenditure, which buys an asset used over several years. Operating costs include salaries, software, rent, professional fees, marketing and logistics. Because opex is a set of standing commitments rather than one-off spend, reducing it means cancelling or reshaping something the business is currently relying on, which makes it slower and more political than it looks on a spreadsheet.
Where should you start with an opex reduction programme?
With a baseline, not a framework. Size the four categories, people, software and cloud, suppliers and outsourced services, and facilities and overhead, before touching any of them. After that comes documenting how the process actually works, spotting the risks and the timesinks, and automating one small step at a time. Applying Lean or Six Sigma to a process nobody has written down produces a diagram rather than a saving.
How much of finance work can be automated?
Roughly 80 to 90% of basic finance tasks such as checking, reconciliation, analysis, forecasting and modelling can be handled by AI. The remaining 10% needs a qualified human who understands the actual challenge and what a good outcome looks like. That split matters for cost cases, because a business case built on removing the last 10% will fail a few months in, usually at the worst possible moment.
Should you cut a finance hire to reduce opex?
It is the most common version of this mistake. Cost pressure frequently stops finance hires happening even where the business case is clear, and the hire being deferred is often the person whose job would be to find the savings. In the Australian market the difference between a compliance-capable hire and someone with real automation, capital raise and acquisition experience is roughly $50k of salary, which is usually less than the cost being carried by not having them.
References
- The sequence I advocate: document how a finance process truly works, then use AI to identify risks and timesinks, before automating one small step at a time. Tom put the end state of that to a CFO who set his function up to run without him in an episode of The CFO Track.
- My read on where automation lands: AI can handle 80-90% of basic finance tasks like checking, reconciliation, analysis, forecasting or modelling, but the last 10% requires a qualified human who understands what the challenge is and what good looks like for the outcome.
- What I see in the market: over the last couple of years significant redundancies, headcount cuts and restructures have led to a dynamic where many highly skilled finance professionals are seeking new roles, but the volume of available jobs has not increased proportionally.
- A pattern across my clients: cost pressures frequently limit hiring for finance roles, even when the business case for new hires is clear.
- Australian market rates as I see them: a half-decent finance hire who can add value beyond compliance might cost $150-160k, but a hire with strong experience in automation, tech, AI, capital raises and acquisitions will be closer to $200k+. A top candidate I know is currently on $230k plus super.
- What employers actually ask me for: candidates who are adaptable and translatable with tech, particularly in automation, process improvement and an interest in AI, without necessarily expecting expert-level AI capability.