Salary benchmarking
Kaan Ersoy2 min read
Who runs this, and what for
| Team | What they decide with it |
|---|---|
| Recruiters and talent teams | Whether an offer is competitive against the companies the candidate is also talking to. |
| Compensation teams | How pay for a title compares across a named peer group of employers, in the states and cities Glassdoor reports. |
The calls, step by step
1. Glassdoor company search
GET /v1/glassdoor/companies
2. Glassdoor salaries
GET /v1/glassdoor/salaries
3. Glassdoor search
GET /v1/glassdoor/jobs
1. Find each employer's id
Company search takes a name and returns employer_id and slug, with the company's rating and how many salaries it has. Both ids are needed for every salary call, and they do not change, so store them.

The page on glassdoor.com that this call reads, as a signed-out visitor sees it. The response below is what Clair returns from it. GET /v1/glassdoor/companies?q=stripe · results.0
{ "position": 1, "employer_id": "671932", "name": "Stripe", "slug": "Stripe", "logo": "https://media.glassdoor.com/sql/671932/stripe-squarelogo-1610580619584.png", "rating": 3.8, "reviews_count": 1588, "salaries_count": 4659, "jobs_count": 4, "url": "https://www.glassdoor.com/Overview/Working-at-Stripe-EI_IE671932.11,17.htm" }Recorded response. Live values differ. 2. Read pay by job title
Salaries returns ten job titles per page, most-reported first, each with base and total pay at the 10th, 25th, 50th, 75th, and 90th percentiles, the number of reports, and when the last one came in. Page until you have the titles you need; the common titles sit in the first few pages. Pass location with a state or city from the response's locations to get pay for that place instead of the whole country.
GET /v1/glassdoor/salaries?employer_id=9079&slug=Google&page=2 · results.0
{ "position": 11, "job_title": "Software Engineer III", "job_title_id": "4514", "salaries_count": 950, "currency": "USD", "period": "year", "base_pay": { "p10": 159325, "p25": 176568, "p50": 197943, "p75": 221906, "p90": 245922 }, "total_pay": { "p10": 271280, "p25": 321338, "p50": 390969, "p75": 492142, "p90": 599160 }, "last_reported_at": "2026-07-24T07:10:52.947Z" }Recorded response. Live values differ. 3. Compare with posted pay
Self-reported pay lags the market. Posted ranges on current openings show what employers offer now: Glassdoor job search returns salary with min, max, and whether the employer or Glassdoor supplied it.
GET /v1/glassdoor/jobs?q=software+engineer&page=1 · results.0.salary
{ "min": 110000, "max": 160000, "median": 135000, "currency": "USD", "period": "year", "source": "employer" }Recorded response. Live values differ.
A script to start from
salary_bands.py
# Salary bands for chosen job titles at the employers you hire against.
# Out: salaries.csv, one row per employer and title, base and total pay.
import csv, os, requests
API = "https://api.clair.im"
HEADERS = {"Authorization": f"Bearer {os.environ['CLAIR_API_KEY']}"}
EMPLOYERS = ["Stripe", "Datadog", "Cloudflare"]
TITLES = {"software engineer", "senior software engineer", "product manager"}
MAX_PAGES = 5 # 10 titles per page, most-reported first
def get(path, **params):
r = requests.get(f"{API}{path}", headers=HEADERS, timeout=90, params=params)
r.raise_for_status()
return r.json()
with open("salaries.csv", "w", newline="") as f:
out = csv.writer(f)
out.writerow(["employer", "title", "reports", "base_p25", "base_p50",
"base_p75", "total_p50", "last_reported_at"])
for name in EMPLOYERS:
employer = get("/v1/glassdoor/companies", q=name)["results"][0]
for page in range(1, MAX_PAGES + 1):
body = get("/v1/glassdoor/salaries", employer_id=employer["employer_id"],
slug=employer["slug"], page=page)
for row in body["results"]:
if row["job_title"].lower() in TITLES:
b, t = row["base_pay"], row["total_pay"]
out.writerow([employer["name"], row["job_title"], row["salaries_count"],
b["p25"], b["p50"], b["p75"], t["p50"],
row["last_reported_at"]])
if not body["next_cursor"]:
break
Set CLAIR_API_KEY to your Clair key; one plan covers the Glassdoor API. Each call spends that API's rate from the plan's monthly balance.
What it costs per month
| Schedule | Requests | Plan | Per month | Per 1,000 |
|---|---|---|---|---|
| 10 employers, refreshed monthly10 company searches plus 5 salary pages each = 60 requests. Cache the ids and later months cost 50. | 60 | Starter | $30 | $500 |
| 50 employers, refreshed monthly50 × (1 search + 5 salary pages) = 300 requests. | 300 | Starter | $30 | $100 |
| 500 employers, 10 pages each, monthly500 × (1 search + 10 salary pages) = 5,500 requests, for a market-wide benchmark. | 5,500 | Starter | $30 | $5.45 |
What to plan for
- Glassdoor salaries are reported by employees and candidates. Titles with few reports (salaries_count) are noisy; set a minimum before you use a number.
- Pay is for the location the response names: the employer's home country by default, or a state or city from locations when you pass location. Only the places Glassdoor lists for that employer are available. Do not mix yearly and hourly rows.
- Titles are not levels. Glassdoor groups by title, so a band for a senior or staff level needs titles that name the level, and there is no filter by title: page through the list to find it.
- Up to 20 pages of 10 titles per employer. Rare titles at large employers may be past that.
- Job titles are Glassdoor's, not yours. Map them to your own levels once, by hand, and keep the mapping.
Common questions
What is the difference between base pay and total pay?
Base is salary. Total pay adds the cash bonus, stock, and profit sharing people reported. At large tech companies total pay can be twice base, so compare like with like.
How current is the data?
Each row has last_reported_at, the date of its most recent report. The percentiles cover reports over a longer window, so they move slowly.
Can I get salaries for a title across all companies?
Not in one call: salaries are read per employer. Read the employers you care about, or use posted ranges from job search for a market-wide view.

Glassdoor API
Crunchbase API
Instagram API
Bloomberg API
Associated Press API