The most Organizations don’t face money problems. They face vision problems.
When an organization grows beyond 80 or 100 employees, we begin to see a correlation between revenue and headcount, but once that threshold is reached, profit margins begin to deteriorate, and do so in a way that cannot be easily attributed to any one factor.
The CFO is tasked with analyzing and reporting why HR costs are disproportionately large. We usually discuss the organizational chart in that meeting. What is rarely considered in this meeting is whether the company actually has the ability to see what they are spending their money on.
The entity is visible to accounting through invoices and payroll, but lacks visibility into the operational layers below the surface. When you have work already done, decisions sit on someone’s desk for a week, hiring takes three weeks because no one is documenting the process, and meetings are held to discuss questions that should be resolved in writing at least six months in advance. Although there is no budget line for that cost layer, it exists, and as the organization continues to grow, so does the cost with the company.
If you spend time with organizations that continue to maintain margins during periods of growth, you’ll find that most dollars leak out of the organization through loopholes in processes long before they reach headcount.
The “re” problem: Where does the margin actually go?
Repetition leads to the elimination of edges. Multiple instances of “redo” indicate broken upstream processes, either due to a bottleneck, a lack of closure in ownership, or a lack of normalized communication that has become completely invisible—one of every “redo” is due to means of losing visibility due to broken upstream processes.
Work again. Reconfirmation. Explanation again. Restart. All of the above uses resources/money without creating anything new. For example; something was done wrong, as a result of which the work was done again; that type of activity will result in further escalation due to the lack of a documented first response; the recruit did not spend enough time to ensure the existing knowledge of the company, while the new employee was there, he rebuilt that information from scratch; a restarted project due to scope changes made after execution was not resolved until too late…etc.
For example, a logistics company with a fairly large presence in its industry tracked very high incidences of shipment errors throughout the entire workflow and discovered that three teams were essentially working on three copies of the same product database. No one had synchronized them in about eight months. After syncing and resolving the issue, their costs to resolve all transactions resulting from these shipping errors (calls to customers for errors, product returns, rush-shipping products, etc.) were over $40,000 per quarter. All shipping error issues were resolved within 45 days of syncing all three product databases.
More often than not, rework remains invisible because no one is seriously looking for it. It seems like overtime, late deadlines, and a persistent feeling that the team is always capable of work, but the results are not commensurate with the efforts. People work. It just happens more than once.
It’s a cost conversation that most companies never have. Expenses are checked in budget reviews. Won’t do it again. It is hidden in execution, treated as a cost of doing business rather than a failure in process design.
The Math on Layoffs Rarely Works All the Way
Eliminating positions reduces payroll costs, has a positive impact on upcoming quarterly reviews, and provides a good presentation to the board. However, the downstream costs of this decision are more difficult to attribute to the original business case.
The cost of losing an employee when an employee leaves and hires a new employee is about $6,000 in lost productivity; in addition, there will be about $3,000 in regards to hiring a new employee. If there are several roles in which this situation occurs, it quickly converges. HR.com’s 2025 Retention Report shows that for every dollar spent on retention, organizations get back nearly $3 in lost recruiting and retraining costs. Additionally, nearly 70% of organizations report increased workloads for remaining employees after significant downsizing decisions, with no potential raises.
Quality will suffer. Workers with other opportunities will look for other jobs. Each time there is an exit, the organization will face additional switching costs. The cycle will repeat over the next 6-18 months. When measurable data shows this behavior, the initial decision will long be questioned.
In addition, there will still be an approved process for employee transitions, an approved bottleneck for employee termination approval, and a paperless process for employee transitions. There will be fewer people acting under more pressure than these broken processes, with no margin for error
Automation works best when the process under it is clean
According to one report, companies that implement structured business process automation report an average rate of return of approximately 240% within the first 6-9 months of implementation. Companies implementing BPA have increased adoption rates from ~20% in 2021 to ~70% by 2025. Companies that report consistent rates of return are doing one thing – they’re fixing the core process first and then building automation on top of it.
Automating the broken approval process does not enhance decision making. Conversely, automating a broken workflow results in a faster transmission of the wrong decision.
A case study is the purchasing team of a large manufacturing company that spent four months developing an automated Purchase Order routing system before realizing that the underlying logic of the approval process contained four (4) additional approval stages that existed long before the creation of the existing financial organization. The automated system worked as designed and supported a lengthy approval process that was never questioned because the company was smaller. After eliminating unnecessary approval processes and re-engineering processes, the time to approve a Purchase Order was approximately 11 days less than before, resulting in increased financial benefits from an additional two months of unblocked/allowed purchasing activity. Software for system automation was a secondary driver of benefits received.
Structured automation also accommodates workload growth in such a way that once the margins are squeezed, headcount additions cannot easily accommodate. Gartner projects that about 69% of day-to-day management tasks could be significantly automated by the end of 2025 – invoice processing, status reporting, approval routing, compliance checks. Recoverable hours are actually a transition towards judgmental work. Global IT spending is forecast to reach $6.15 trillion in 2026, with generative artificial intelligence spending expected to grow 80.8% that year, most of which is focused on the everyday decision-making category. Companies deploying that infrastructure now are solving the cost problem later, with less time and fewer options.
What operationally intensive companies actually do
Process debt behaves like technical debt – it quietly accumulates interest and compounds until someone is forced to deal with it at the worst possible moment. A workflow built for 40 people out of 200 that no one reviews costs time and clarity every week. Rather than waiting for a crisis to create urgency, operationally tight companies are testing processes on a schedule. Less dramatic. It is significantly cheaper.
Escalation patterns are another place where costs pile up without appearing on any budget line. When the same questions continue to climb into senior management, the reflex is to question the skills in the junior ranks. More often than not, it’s a policy loophole that was never properly documented. Seniors’ time is precious. Slow decisions have their costs. And over the months, the culture of constant promotion quietly pushes away the capable people who finally stop asking and start walking away.
Onboarding time is probably the least discussed indicator of operational health, which is a mistake. Short onboarding reflects the quality of documentation, process clarity, and institutional knowledge that resides in systems rather than individuals. When a key person leaves and the organization absorbs it without visible disruption, it’s years of thoughtful infrastructure doing its job—the same infrastructure that keeps day-to-day costs from spiraling upward.
The proliferation of tools compounds this in a way that most companies don’t appreciate. A mid-sized business pays 20-40% more for SaaS space than it actually uses. When growth was fast and cash was available, no one was checking subscriptions closely. Now that the budget is quietly funding seven overlapping tools that do versions of the same job, the team has found solutions for half of it. The costs sit there, spread out enough that no line item is relevant, so large that they add up over the course of a year.
None of these fixes require a transformation program or an outside consultant. They require someone empowered enough to treat operational health as a serious ongoing discipline, not a cleanup project that happens when something goes wrong badly enough to warrant attention.
Dismissal is a diagnosis, not a cure
When a company turns to downsizing to solve a margin problem, it usually solves the symptom with the cleanest line item. Salary is visible. Action produces measurable short-term results. It means determination. What it doesn’t do is address the bottom line—processes that don’t scale, decisions move too slowly, work is duplicated among teams that quietly lose sight of what each other is doing.
Some companies learn this after a longer period of layoffs, rehiring, and morale to recover from the initial margin problem. Establishing a vision first, refining it, and coming to headcount decisions from an informed position are more difficult to present at a board meeting. Results last longer.
Today, cuts are cutting costs. Systems determine if it will come back tomorrow.
