How to build salary bands for an international sales office

To build salary bands for an international sales office, you need a compensation benchmarking methodology that works before you employ anyone in the country. This guide sets one out for a company opening a sales office in the United Kingdom: map each hire to a sales role, read the quartiles of advertised base pay and on-target earnings, test the evidence behind them, convert on one dated basis, and set a midpoint and a spread. It ends with a cross-check against official UK earnings statistics.

How to build salary bands for an international office: the benchmarking methodology

Build each band from the market's advertised pay for one sales role in one country. Take the quartiles of base pay and on-target earnings, confirm that enough postings and employers stand behind them, convert to the local currency at one dated rate, place the midpoint where your pay policy sits, and set the band's width around it.

Throughout, pay is what an employer advertises for an open role, an offer to a new hire, and every converted figure carries its rate and the rate's date so that anyone can redo it.

Using job posting salary data to set international pay bands

Job postings show the pay employers offer for a role in a country, and a sales posting may state base pay and on-target earnings apart. Group postings by sales role and country, take quartiles rather than averages, record how many postings and employers stand behind each, and read the result as the market's asking price for new hires.

Sales Pay Benchmarks, part of the hiring dataset, does that grouping every week. It has one row per closed week, sales role and country, with the quartiles of advertised base pay and on-target earnings on a yearly US-dollar scale and the number of postings behind them. A group is written only when at least five postings stand behind it, and the country ALL pools every country. The guide to benchmarking sales compensation sets out the columns and places an existing plan against them; this guide builds a new band from them.

Start by mapping each planned hire to a role of the sales labels, which the data dictionary defines.

Then read a run of recent weeks for the office's country and for ALL, side by side.

select week_start, role, country, postings,
       base_p25_usd, base_median_usd, base_p75_usd,
       ote_p25_usd,  ote_median_usd,  ote_p75_usd
from sales_pay_benchmarks
where role in ('sdr_bdr', 'account_executive', 'sales_engineer')
  and country in (:country, 'ALL')
  and week_start >= :first_week
order by role, country, week_start;

Here :country is the two-letter ISO 3166-1 code of the office's country, which you look up, and :first_week is a Monday some weeks back. A posting sits in the group of every week it stays open, so a run of weeks shows whether the level is stable rather than how much evidence there is. A week with no row for the country had fewer than five postings in that group.

Checking the evidence behind each quartile

The postings count in Sales Pay Benchmarks covers postings that state a base or on-target earnings, so the two sets of quartiles can rest on different postings, and a count of postings is not a count of employers. To see both, pool the postings open at any time in a window from Job Postings, keep one row per posting and count. The query assumes JSON cells are loaded as jsonb and empty cells as null.

with latest as (
  select distinct on (posting_id) *
  from job_postings
  order by posting_id, last_seen_at desc
),
pool as (
  select * from latest
  where sales_labels ->> 'role' = 'account_executive'
    and country = :country
    and employment_type = 'full_time'
    and first_seen_at < :window_end
    and (closed_at is null or closed_at >= :window_start)
)
select
  count(*)                                                       as postings,
  count(distinct company_id)                                     as companies,
  (select max(n) from (select count(*) as n from pool group by company_id) c) as top_company_postings,
  count(*) filter (where sales_facts ->> 'ote_usd' is not null)  as with_ote,
  count(*) filter (where sales_facts ->> 'base_usd' is not null) as with_base,
  count(*) filter (where sales_facts -> 'ote' ->> 'currency' = :local_currency) as ote_in_local_currency,
  percentile_cont(0.5) within group (order by (sales_facts ->> 'ote_usd')::numeric)    as ote_median_usd,
  percentile_cont(0.5) within group (order by (sales_facts ->> 'base_share')::numeric) as base_share_median
from pool;

Read four things before any figure goes into a band.

• Evidence per figure. with_ote and with_base count the postings behind each pay measure separately.
• Employers. companies and top_company_postings show whether the pool describes a market or one employer's pay grid.
• Currency. ote_in_local_currency counts postings whose on-target earnings are written in the office's currency (:local_currency, as the data writes it, read from a sample first).
• Split. base_share_median is the median of the base share each posting states. The ratio of the base median to the on-target median comes from two sets of postings that need not match, and is a different figure.

Write your floors down before you run the query: a minimum of postings, a minimum of employers and a ceiling on one employer's share of the pool. Filter on labels ->> 'seniority' for a band per seniority level, or on sales_labels ->> 'segment' for an office that sells to one segment.

Converting to the office's currency on one stated basis

Sales Pay Benchmarks and the US-dollar fields of Job Postings are already yearly, so the pay period is settled and only the currency is left. Choose one source of rates and one date or averaging period, apply it to every figure in the band, and print both beside it.

• European Central Bank. Its euro foreign exchange reference rates quote each currency, the US dollar and sterling among them, against the euro. The page, read on 7 October 2026, says they are usually updated at around 16:00 CET on every working day except TARGET closing days and are published for information only, with their use for transactions strongly discouraged. Sterling against the dollar is therefore a cross rate of two euro rates.
• Bank of England. It publishes daily spot rates against sterling. Its notes on the spot rate data, last updated 11 March 2025, describe a spot rate as one for settlement in two days, list monthly, quarterly and annual averages and end-of-period rates beside the daily ones, and state that the figures are statistical data and not official rates. An average over a month or a year keeps one day's move out of a band you will hold for a year.

Then find the rate already inside the data. Job Postings carries pay both as the employer states it and on the yearly US-dollar scale, so for yearly pay stated in the local currency the ratio of the two is the rate that was applied.

select
  date_trunc('month', first_seen_at)                 as month_first_seen,
  count(*)                                           as postings,
  min(salary_usd_annual_min::numeric / salary_min)   as lowest_usd_per_unit,
  max(salary_usd_annual_min::numeric / salary_min)   as highest_usd_per_unit
from job_postings
where country = :country
  and salary_currency = :local_currency
  and salary_period = 'year'
  and salary_min > 0
  and salary_usd_annual_min is not null
group by 1
order by 1;

If the lowest and highest values agree across months, one rate stands behind the scale; if they differ, the scale mixes rates. Where one rate stands behind it, converting the US-dollar quartiles back at that rate returns the figures employers wrote, where they wrote in the local currency. Converting at any other rate moves the whole band by the gap between the two rates with no change in the market, so record which basis you used.

Setting the midpoint and the spread

The market position sets the midpoint: at the median to match the advertised market, or between the median and the upper quartile to lead it. The width then sets the minimum and maximum, and the market's spread between the quartiles is the measure to hold it against.

minimum       = midpoint x (1 - h)
maximum       = midpoint x (1 + h)
range spread  = (maximum - minimum) / minimum = 2h / (1 - h)
market spread = (p75 - p25) / p25

A sales band is two linked bands. Set the on-target earnings band first, take the base band from the split, and the variable pay is the difference between them at each point. A range spread well below the market spread leaves little room for experience within one band; well above it, the band overlaps the level above or below. Repeat the steps for each role, since each has its own split and spread.

What official wage statistics add, and how they differ

Official earnings statistics measure what employees are paid, where job postings measure what employers offer. In the United Kingdom this guide uses the Annual Survey of Hours and Earnings (ASHE) of the Office for National Statistics, and each point below comes from its own pages, read on 7 October 2026.

• What it measures. The 2025 results, released on 23 October 2025, rest on employer responses for a 1% sample of employees identified through HM Revenue and Customs PAYE records, with an achieved sample of 174,000. Every 2025 estimate is provisional and refers to the pay period that includes 30 April 2025, and the 2024 estimates have since been revised.
• Who is counted. Annual pay figures cover employees who have held their job for a year or more. Full-time means more than 30 paid hours a week, or 25 or more in the teaching professions, and the median is the statistics office's preferred measure of average earnings.
• What pay includes. The ONS glossary of income and earnings, last revised 17 December 2019, defines gross earnings as earnings before tax deductions, National Insurance contributions among them, and incentive pay as what an employee is paid for meeting a performance or productivity objective, commission and bonus included. The ASHE methodology and guidance, released 18 January 2023, lists gross annual pay and annual incentive pay among the variables available.
• How jobs are grouped. By occupation under the Standard Occupational Classification 2020, not by sales role. ASHE Table 14, released 23 October 2025, gives annual estimates of paid hours and earnings for UK employees by sex, full-time and part-time, and four-digit occupation. In SOC 2020, unit group 3556 is Sales accounts and business development managers, which took in part of unit group 3545 of SOC 2010. Mapping each sales role to a unit group is your judgement.

The posting quartiles price a new offer. ASHE shows what people in the matching occupation are paid, and its annual incentive pay against gross annual pay gives a realised variable share to set beside the advertised split. The ONS has also released salary information of online job adverts at UK level, January 2017 to April 2025, as an ad hoc table on 20 May 2025.

Pay transparency and job postings as a salary benchmark for an international sales office

Job postings make a sound salary benchmark for a sales office in a country where enough employers state pay for the role, and a rule that requires pay in the advert adds to that number where it applies. Check the postings and employers behind each quartile in the office's country first, and read the all-countries row beside it.

Rules differ on where pay must appear. Article 5(1) of Directive (EU) 2023/970 gives applicants a right to information on the initial pay or its range, given for example in a published vacancy notice, before the interview or in another way, and the European Commission states that Member States must transpose the directive into national law by 7 June 2026, so read the national law of the office's country for what a posting there must carry. Because the information can reach the applicant after the advert, a posting in a Member State can still carry no pay. In the United Kingdom, the Government Equalities Office announced on 8 March 2022 a pay transparency pilot in which participating employers list salary details on job adverts and stop asking about salary history.

Test the depth country by country: run the evidence query for each country you are weighing and divide with_ote by postings. Tracking sales team build-out shows which companies hire sellers in a country and when one opens its first sales posting there.

Worked example: an illustrative UK band for Acme Robotics

Acme Robotics is an invented company, and every figure below is invented for the arithmetic. None is a Fokals figure or a market finding.

Assumptions:

• Acme opens a sales office in the United Kingdom and hires full-time account executives.
• The benchmark for the role in the country: on-target earnings of USD 112,000 at p25, USD 128,000 at the median and USD 148,000 at p75; base pay of USD 62,000, USD 70,000 and USD 80,000.
• The pooled window holds 64 postings from 23 employers, the largest with 9. 58 state on-target earnings and 61 state a base, and the median base share is 0.55.
• Acme's floors: 30 postings, 10 employers, and no employer above a quarter of the pool.
• Policy: the on-target midpoint at the median, and h = 0.15, a band from 85% to 115% of the midpoint.
• The rate behind the US-dollar scale, read with the test query: 1 GBP = 1.28 USD. A later rate for the sensitivity check: 1.32. Both are assumed, not market rates.

Working:

1. Evidence: 64 postings against a floor of 30, 23 employers against 10, and 9 of 64 postings, 14.1%, from the largest employer. The on-target figures rest on 58 postings. The group passes.
2. Midpoints: on-target earnings USD 128,000. Base 0.55 x 128,000 = USD 70,400, which sits between the base median and p75. Variable pay at the midpoint: 128,000 - 70,400 = USD 57,600.
3. On-target band: 0.85 x 128,000 = USD 108,800 to 1.15 x 128,000 = USD 147,200. Range spread 147,200 / 108,800 - 1 = 35.3%, against a market spread of (148,000 - 112,000) / 112,000 = 32.1%. The band starts USD 3,200 below p25 and ends USD 800 below p75.
4. Base band: 0.85 x 70,400 = USD 59,840 to 1.15 x 70,400 = USD 80,960.
5. Conversion at 1.28, each US-dollar figure divided by the rate:

Variable pay at the midpoint is GBP 45,000, which is 57,600 / 1.28. At the later rate of 1.32 the on-target midpoint would be 128,000 / 1.32 = GBP 96,970, about 3.0% lower, while UK employers advertise exactly what they did. Acme therefore prints the basis beside the band: converted at the rate behind the US-dollar scale, 1.28. Last, it maps the role to the SOC 2020 unit group it judges closest and reads the ASHE figures for that group, expecting a gap for the reasons in the table above.

Frequently asked questions

The queries and code on this page are examples to adapt. Test them in your own environment before you rely on them.

What this page says about the products it names was checked against their public documentation on 7 October 2026. Product and company names are trademarks of their owners. Fokals is not affiliated with them or endorsed by them.