The easiest remote-versus-office comparison fits into two lines: salary and rent. It is also the comparison that leaves the most out.
The issue is not that remote work contains mysterious “hidden costs.” Some costs simply change form. What used to happen by walking to another desk may become documentation, meetings, software, waiting time, or a more deliberate onboarding process.
That does not make remote work worse. It means the useful comparison is between two complete operating systems, not two salaries.
First define the alternative
“Remote versus office” sounds clear until you try to model it. Is the real alternative five days in an office? Hybrid work? Hiring locally versus expanding the search to another city or country? Keeping a role in-house versus using a provider?
The baseline matters. A company that already pays for office space, software, and equipment does not automatically save all of those costs because two people work from home. A company deciding whether to open an entirely new location may avoid a much larger block of expense.
Five lines that belong in the model
| Cost line | What to measure |
|---|---|
| Total compensation | Salary, payroll costs, benefits, and recruiting costs. |
| Physical space | Rent, utilities, furniture, maintenance, security, and actual seat usage. |
| Tools | Additional software, duplicate licenses, hardware, and support. |
| Coordination | Meeting hours, time-zone delays, rework, and documentation overhead. |
| Turnover and onboarding | Time to fill roles, ramp-up, and lost productivity during transitions. |
The key is to measure the difference between scenarios. If the company pays for the same collaboration suite and laptop either way, do not assign the full cost to remote work. Count what changes because of the decision.
Coordination is the hardest cost to see
A meeting does not cost its duration. It costs the duration multiplied by the number of people in it, plus the disruption around it and, sometimes, the delay caused by waiting across time zones.
That makes a remote hire with several overlapping work hours different from one where nearly everything must move asynchronously.
Latin America can be attractive to North American employers partly because of time-zone proximity, not only compensation differences. But that should not be turned into one universal salary figure. Colombia, Mexico, Argentina, and Brazil are different labor markets, and a senior engineer does not follow the same economics as a junior analyst.
An example—treated as an example
Suppose a company compares two ways to fill one role. In Scenario A, total annual compensation plus the physical workstation equals 100 units. In Scenario B, remote hiring lowers compensation and removes part of the space cost, reducing direct cost to 72.
If Scenario B also adds 4 units of software and support, 5 of coordination, and 3 of onboarding/documentation, the savings are not 28. They are 16.
This is not a claim that remote work “saves 16%.” It illustrates a more useful point: headline savings and net savings are different numbers.
A formula that avoids double counting
Net scenario cost = total compensation + incremental physical-space cost + incremental tools + estimated coordination + expected onboarding/turnover cost.
The word “incremental” matters. If a cost exists in both scenarios, it does not decide the comparison.
For coordination, one workable approximation is additional meeting or handoff hours multiplied by the loaded labor cost of the people involved. It is not perfect, but it prevents meetings from being treated as free.
What the spreadsheet will not capture cleanly
Some factors eventually affect money but are difficult to price in advance: access to talent, hiring speed, informal learning, culture, concentration, relocation needs, and resilience to local disruptions.
Instead of inventing a precise dollar value, declare them as assumptions and revisit them. “We expect a broader search to reduce time-to-hire” is a testable hypothesis. “Remote increases productivity by 23%” with no defensible source is decoration.
Make the model learn from reality
After six or twelve months, compare the budget with what actually happened. Did software spending rise? Did meeting hours grow? Did hiring time fall? Was turnover different? Did the company really shrink office space, or did it keep paying for an underused building?
That review turns a one-time cost comparison into a learning system. Without it, any model—for or against remote work—eventually becomes a set of assumptions wearing spreadsheet formatting.
The useful conclusion is not that remote work is always cheaper or that offices are always more efficient. It is that cost depends on what genuinely changes when one operating model replaces another.
