This is an interesting question, and does require some thought. To which level of role should geo-differentials be applied?
To start with, what is a geographic differential? Well, typically when creating salary ranges, or benchmarks, rather than creating a midpoint for every position/level in every city in a given country, we will group cities and regions together and treat them the same for the purposes of compensation. For example, you may decide that the Bay Area should qualify as a single location given it’s significant pay premium, but that Seattle, LA, and San Diego all pay at comparable levels so can be grouped together.
This is great for building national compensation structures but it does get a little more complicated when thinking about senior levels.
Let’s take a CEO. A CEO will not usually be paid differently based on their location since cost-of-living is easily covered by base salary and senior executives will typically be recruited from a national talent pool. Going a step further, it will often be the case that executives are benchmarked internationally so allowing for regional pay differentials doesn’t make much sense at all.
However if we take a support-level individual such as an administrative assistant or assembler, it is unlikely that a company would be willing to pay the recruiting search and relocation cost to bring someone across the country so the talent pool is usually local.
But what about those senior positions in-between? That is where things get a little tricky. Are Directors recruited from a national talent pool or local? What about VPs or SVPs?
In reality it is a bit of both. You probably will have to pay more for the Director in the Bay Area than you will for the Director in Utah but at this level, the pool of talent becomes shallower and most companies will pursue talent, at least to some extent, on a national basis. While difficult to fully validate for all use cases, this pattern is generally reflected in market data with regional differences in pay being more pronounced at the junior end of the organization, and increasingly less pronounced higher up.
One solution is to ignore this phenomenon and set differentials equally across all levels. In smaller companies, in particular, this is perfectly fine and if there are nuances in pay, they can easily be accommodated by the ranges. An alternative is to scale the geo-differentials based on seniority so that midpoints gradually converge towards the top of the organization:
So bear in mind that pay ranges are there for a reason and it is easy to over-engineer salary midpoints to the point of “false accuracy”. But there are alternative approaches that can be used and plenty of ways to skin this particular cat.
“Breaking rules isn’t interesting. It’s making up new ones that keeps things exciting.” ~ Christopher Nolan