Geographic Pay Differentials: How to Build Pay Zones

Geographic Pay Differentials: How to Build Pay Zones

Someone on your team just asked to move from Denver to Boise. Someone else moved to Austin eight months ago and never told anyone. A third person is in the Bay Area, hired remote, on a range you built for the national market. Every one of those is a pay decision, and right now you are probably making each one by hand.

Geographic pay differentials replace that guessing with a rule: pick a reference market, group your other locations into a few zones, and apply a percentage to the structure — not to the person. This guide covers how to build zones from scratch, how to set the percentages, and how to handle the remote-employee questions that break most policies in the first month.

TL;DR

• Geographic differentials adjust the structure for cost of labor, not the person for cost of living. Different data sets, different answers.
• Anchor on a control point — your largest market or a national 50th percentile — then express every other zone as a percentage of it.
• Three to five zones covers most US employers. More than six and you are maintaining a spreadsheet nobody trusts.
• Set the geographic rule after job evaluation, never instead of it. Internal level comes first; location modifies the range.
• Write the placement rule down before you need it: which address governs, what happens when someone moves, and whether pay goes down.

Cost of labor is not cost of living

This distinction decides whether your zones survive scrutiny.

Cost of living measures what it costs an individual to exist somewhere — housing, groceries, taxes, transport. It is a consumer statistic, and it varies enormously with personal choice.

Cost of labor measures what employers in that market actually pay for a given job. It is a market statistic drawn from survey data, and it is the only one that tells you whether you will win the candidate.

The two correlate, but not tightly, and they diverge in exactly the places you care about. A market can be expensive to live in and merely average to hire in — university towns and resort areas do this routinely. Build zones on cost of living and you will overpay in some markets and lose every offer in others, with no way to explain the pattern.

WorldatWork's research on geographic pay policies found that cost of labor is overwhelmingly the greater influence on how organizations set their approach, and that more than half of employers (56%) use city or metro area as the geographic indicator (WorldatWork). Use cost of labor. Keep cost of living for relocation conversations, where it belongs.

Step 1: Pick your control point

Every zone structure needs a reference — the market whose rates equal 1.00. Three reasonable choices:

Your largest employee population. If 40% of your people sit in one metro, make that metro 1.00. Most of the organization then needs no adjustment at all.

A national 50th percentile. Build the base structure on national survey data and treat every location as a deviation. This fits a distributed or remote-first company with no dominant hub.

Your highest-cost market. Set San Francisco or New York at 1.00 and discount everywhere else. Administratively clean but psychologically awful — nobody enjoys being told they are on a discount.

Take the first or second. Whichever you pick, write it down, because the control point is what people forget in eighteen months when someone asks why Zone 2 is 0.94.

Step 2: Build the zones from market data, not from a map

Pull cost-of-labor data for 15 to 25 benchmark jobs spanning your levels and functions — an engineer, a finance analyst, a warehouse supervisor, a support rep. Functions differ in geographic sensitivity, and using only tech roles will exaggerate the spread.

For each metro where you employ people, calculate the median differential across that benchmark set against your control point. You now have raw percentages — Denver 0.96, Austin 0.98, Boise 0.89, Raleigh 0.94, San Jose 1.19. Now group them. Do not create a zone per city.

Four zones handles the large majority of US footprints. Resist adding a fifth until a market sits genuinely 20% off the nearest zone with enough people in it to matter — every extra zone multiplies the ranges you maintain, the survey cuts you buy, and the exceptions you argue about. A zone is a range of markets deliberately treated the same, which is what stops you relitigating a 2% survey wobble every year.

Step 3: Apply the differential to the structure, not the person

This is where geographic pay goes wrong most often. Your salary structure has grades, each with a minimum, midpoint, and maximum. The differential multiplies those three numbers to produce a zone-specific range. It does not multiply anyone's salary.

So if Grade 12 has a midpoint of $118,000 at control:

• Zone 1 (1.20): midpoint $141,600
• Zone 3 (1.00): midpoint $118,000
• Zone 4 (0.92): midpoint $108,560

An individual's pay is then positioned within their zone's range using the normal logic — performance, experience, compa-ratio targets. Two people in the same grade in different zones can have very different salaries and identical compa-ratios. That is the system working correctly.

The ordering keeps location out of the job-worth decision entirely. The grade comes from what the job requires — the skill, effort, responsibility, and working conditions captured in your point-factor evaluation. A senior analyst in Boise does the same job as a senior analyst in Boston and lands in the same grade. Only the range changes. Let geography influence the grade itself and you have built a system where the same work is worth less in some places — very hard to defend to a regulator or a works council.

If your grades came from job titles rather than job content, fix that before you layer zones on top — PointFactors scores jobs against weighted compensable factors so the level is defensible in every market.

Step 4: Decide which address governs

Here is the question that lands in your inbox within a week of launch: whose location counts? You need a written answer for each of these:

• Office-based and hybrid employees. Almost always the assigned work location. WorldatWork found employers most commonly use nearest work location (45%) or reporting location (31%) here.
• Fully remote employees. Almost always the employee's residence. More than half of employers tie full-time remote workers to where they live.
• Employees who split time. Pick the primary residence and require a threshold — say, more than half the year — before a change applies.
• Employees who move without telling you. This will happen. Your policy needs a notification requirement and a stated consequence, or you will find someone has been in the wrong zone for two years with no clean way to fix it.

The federal government offers a reference model. OPM sets locality rates by official duty station, not home address, across 58 defined locality areas, and the spread is substantial: the San Jose–San Francisco–Oakland area carries a 46.34% locality payment on top of base General Schedule rates for 2026 (OPM). The duty-station rule is unambiguous and cheap to administer. It also produces results employees sometimes find unfair. That is the trade-off either way.

Step 5: Write the movement rule before you need it

The hardest question in geographic pay is whether pay goes down when someone moves to a cheaper market. There is no universally correct answer, but there is a correct process: decide in advance, in writing, and apply it consistently. Three common approaches:

Full adjustment. Pay moves to the new zone immediately, up or down. Cleanest internally, hardest on the individual, and it discourages moves you might otherwise be happy to allow.

Downward protection with a freeze. Pay does not drop, but the employee receives no increases until the zone range catches up. This is the most common compromise — it avoids a visible cut while resolving the gap over two to four years. It also creates red-circled employees you need to track deliberately.

No adjustment on voluntary moves. Pay stays with the original zone. Simple and popular, but it seeds internal inequity — two people doing the same job in the same city on different structures, for reasons neither can see. Over enough moves that becomes a real pay equity problem.

Whichever you choose, apply upward adjustments on the same timeline as downward ones. A policy that raises pay immediately for a move to San Francisco but protects it indefinitely for a move to Omaha is not a policy, it is a slow budget leak.

Data quality and the four ways this fails

Your zone percentages are only as good as the survey data underneath them.

Match jobs on content, not title. "Analyst" spans four grades in most survey libraries. Use the job's evaluated level and scope, the same discipline that governs salary benchmarking generally. Where a metro has fewer than about 20 observations, blend sources or fall back to the state cut.

Sanity-check against public data. BLS Occupational Employment and Wage Statistics is a free, large-sample reference for how wide the real spread is — software developers showed national employment of 1,687,890 and an annual mean wage of $148,100 in the May 2025 OEWS release (BLS). If a commercial survey says something dramatically different, find out why before building a zone on it.

Then watch for the four predictable failures. Too many zones: nine tiers for a 400-person company means nobody can explain the structure. Zones as a hiring workaround: if a market keeps producing failed offers, the problem is usually the grade or the range width — fix the job architecture instead. Silent drift: people relocate, nobody updates the record, and your assignments stop describing reality; build an annual location attestation into the merit cycle. Compression across zones: raise Zone 4 ranges 4% and Zone 1 ranges 2% and you have narrowed the gap between two populations who talk to each other, so watch pay compression between zones as well as within them.

Frequently asked questions

What are geographic pay differentials?

They are percentage adjustments applied to a salary structure to reflect differences in the cost of labor between markets. A job graded at midpoint $100,000 in the control market might carry a $120,000 midpoint in a premium zone and a $92,000 midpoint in a lower-cost zone. The job's evaluated level does not change — only the range attached to it.

Should differentials be based on cost of living or cost of labor?

Cost of labor. Cost of living measures an individual's expenses and varies with personal choices; cost of labor measures what employers in that market actually pay for the work. Cost of labor determines whether your offers get accepted, and it is what survey data supports.

How many pay zones should we have?

Three to five for most US employers. Fewer than three usually fails to capture real market differences; more than six creates a maintenance burden and an explanation problem. Start with four and split a zone only when a market sits roughly 10 percentage points off its assigned tier with enough employees affected to matter.

Can we cut someone's pay when they move to a cheaper market?

This varies by jurisdiction, generally requires advance written notice, and in some states cannot be applied retroactively. Practically, most employers freeze rather than cut — the employee keeps their salary and receives no increases until the new zone's range catches up. Get it in the policy before the first request arrives, and check state notice requirements with counsel.

Do geographic differentials create pay equity risk?

They can, if applied inconsistently. The risk is not the differential but the exceptions: if senior people negotiate their way out of downward adjustments and junior people do not, the pattern shows up in your next pay equity analysis. Apply zones by rule, document every exception, and audit exceptions by demographic group annually.

How do differentials interact with pay transparency requirements?

Posted ranges generally need to reflect the location where the role will be performed. Zone-based structures make this easier — you already have a documented range per location instead of assembling one per posting. For remote roles spanning multiple zones, most employers post the widest applicable range and note that pay depends on location.

Should executives be zoned?

Usually not. Executive markets are national rather than local. Most employers stop applying zones around the director or VP level.

Geographic pay is one of the few compensation problems where the mechanics are simple and the discipline is everything. Four zones, one control point, a written placement rule, and an annual refresh will outperform a nine-tier model nobody maintains. The part that has to be right first is the grade — a differential applied to a badly leveled job just makes the error more expensive in more places.

If your ranges are built on job titles rather than evaluated job content, zones will amplify that problem rather than solve it. Book a demo and see how point-factor scoring produces grades you can defend in every zone you operate.

Justin Hampton is founder and CEO of PointFactors.