Research Company Reviews on Glassdoor at Scale (2026)

Published April 27, 2026 · 1527 words · For researchers

Thirdwatch's Glassdoor Scraper returns Glassdoor company reviews at $0.008 per record — company_rating, pros (positive theme array), cons (negative theme array), URL — when set to scrapeType: reviews. Built for employer-brand researchers, competitive-intelligence functions, equity analysts watching workforce sentiment, and HR competitive-intelligence groups who need machine-readable review data without manual page-by-page browsing.

Why research company reviews on Glassdoor

Glassdoor remains the largest crowd-sourced US/UK employee-review database. According to Glassdoor's 2024 transparency report, the platform aggregates employee reviews across 1.4 million companies — making it the canonical input for any employer-brand or competitive-culture research targeting US/UK markets. The blocker for systematic access: Glassdoor retired its public partner API in 2018, and competitor research SaaS (LinkedIn Talent Insights, BuiltIn Pro) bundles Glassdoor data behind enterprise licensing.

The job-to-be-done is structured. An employer-brand consultant building a benchmark report wants pros/cons across 30 sector peers. An equity-research analyst monitors workforce-sentiment shifts at portfolio companies as a leading earnings indicator. An HR competitive-intelligence team tracks competitor culture trajectories over time to inform talent-attraction messaging. A trust-and-safety team monitors brand mentions in employee-review surfaces as part of crisis-detection workflows. All reduce to scrapeType=reviews + company watchlist + country → structured review rows.

How does this compare to the alternatives?

Three options for getting Glassdoor review data into a research pipeline:

Paid employer-brand SaaS includes Glassdoor with proprietary dashboards but is priced for full marketing-research departments. The Glassdoor Scraper actor page gives you the data layer at pay-per-result pricing.

How to research company reviews in 4 steps

Step 1: How do I authenticate against Apify?

Sign in at apify.com (free tier, no credit card), open Settings → Integrations, and copy your personal API token. Every example below assumes the token is in APIFY_TOKEN:

export APIFY_TOKEN="apify_api_xxxxxxxxxxxxxxxx"

Step 2: How do I pull reviews for a competitor watchlist?

Set scrapeType: "reviews" and pass company names as the queries.

import os, requests, pandas as pd

ACTOR = "thirdwatch~glassdoor-scraper"
TOKEN = os.environ["APIFY_TOKEN"]

PEER_SET = ["google", "meta", "openai", "anthropic", "stripe",
            "airbnb", "uber", "lyft", "doordash", "instacart"]

resp = requests.post(
    f"https://api.apify.com/v2/acts/{ACTOR}/run-sync-get-dataset-items",
    params={"token": TOKEN},
    json={
        "queries": PEER_SET,
        "scrapeType": "reviews",
        "country": "us",
        "maxResults": 100,
    },
    timeout=3600,
)
df = pd.DataFrame(resp.json())
print(f"{len(df)} review records across {df.company_name.nunique()} companies")

10 companies × 10 review records = ~100 records, costing $0.80.

Step 3: How do I aggregate themes across the peer set?

Group pros and cons by frequency to surface industry-wide vs company-specific patterns.

from collections import Counter

def explode_themes(df, col):
    rows = []
    for _, row in df.iterrows():
        themes = row.get(col) or []
        if isinstance(themes, list):
            for theme in themes:
                rows.append({"company": row["company_name"], "theme": theme.lower()})
    return pd.DataFrame(rows)

pros = explode_themes(df, "pros")
cons = explode_themes(df, "cons")

# Industry-wide themes (mentioned by 5+ companies)
industry_pros = (
    pros.groupby("theme")["company"].nunique()
    .sort_values(ascending=False).head(15)
)
industry_cons = (
    cons.groupby("theme")["company"].nunique()
    .sort_values(ascending=False).head(15)
)
print("--- Top shared positive themes ---")
print(industry_pros)
print("--- Top shared negative themes ---")
print(industry_cons)

Common pros across the set are table-stakes (most tech companies cite "smart colleagues", "competitive comp", "remote flexibility"); common cons are sector-wide pain points (layoffs, slow promotion cycles, work-life balance pressure during crunches). Themes mentioned by only 1-2 companies are the genuine company-specific signals.

Step 4: How do I detect rating drift over time?

Persist daily snapshots and diff company_rating across days for trend tracking.

import datetime, json, pathlib, glob

today = datetime.date.today().isoformat()
snapshot_dir = pathlib.Path(f"snapshots/glassdoor-{today}")
snapshot_dir.mkdir(parents=True, exist_ok=True)
df.to_json(snapshot_dir / "reviews.json", orient="records")

frames = []
for d in sorted(glob.glob("snapshots/glassdoor-*")):
    date = pathlib.Path(d).name.replace("glassdoor-", "")
    snap = pd.read_json(f"{d}/reviews.json")
    snap["date"] = pd.to_datetime(date)
    frames.append(snap)
history = pd.concat(frames, ignore_index=True)

ratings = history.groupby(["company_name", "date"])["company_rating"].first().unstack(level=1)
dates = sorted(ratings.columns)
if len(dates) >= 30:
    ratings["delta_30d"] = ratings[dates[-1]] - ratings[dates[-30]]
    movers = ratings[ratings.delta_30d.abs() >= 0.2].sort_values("delta_30d")
    print("--- Rating movers (30-day) ---")
    print(movers[[dates[-30], dates[-1], "delta_30d"]])

A 0.2+ point drop in 30 days is a meaningful trend signal — typically corresponds to a layoff, leadership change, or major PR event. Cross-validate by manually reviewing a sample of recent reviews on the affected companies.

Sample output

A single review record for one company looks like this. Five rows weigh ~3 KB.

{
  "company_name": "Microsoft",
  "company_rating": 4.2,
  "pros": ["Great work-life balance", "Strong benefits", "Excellent training programs"],
  "cons": ["Slow promotion cycles", "Bureaucratic at scale"]
}

pros and cons are theme arrays — Glassdoor's clustering of common phrases across reviews, not individual review excerpts. For most employer-brand and competitive-intelligence workflows, the theme summary is the right level of detail; individual review-text analysis adds noise without proportional signal. company_rating is Glassdoor's headline 1-5 score, the canonical trend metric for rating-drift tracking.

Common pitfalls

Three things go wrong in production Glassdoor review pipelines. Theme-clustering drift — Glassdoor occasionally re-clusters its theme algorithm, which changes pros/cons array contents without an underlying review change; treat sudden theme-array turnover as a methodology artifact, not real culture shift. Public-vs-private bias — public companies have orders of magnitude more reviews than private companies, so review-volume comparisons across the public/private divide aren't apples-to-apples. Country-localised company-name matching — meta returns Meta Platforms in the US but may return a different local company in non-US domains; always inspect the resolved company URL to confirm you got the intended company before downstream analysis.

Thirdwatch's actor uses Camoufox stealth-browser bypass for Glassdoor's Cloudflare protection — production-tested at sustained weekly volumes. The 4096 MB max memory and 3,600-second timeout headroom mean even 100-company batch runs complete cleanly. Pair Glassdoor with our AmbitionBox Scraper for India coverage where AmbitionBox has materially deeper review data. A fourth subtle issue worth flagging: Glassdoor occasionally hides reviews on companies that file legal complaints under their Counterclaim Notice program, which means a sudden drop in pros/cons array length without rating change usually indicates content-moderation activity rather than real culture shift. A fifth pattern worth knowing for crisis-monitoring: Glassdoor caps the rate at which a single company's rating can move publicly (the displayed score is a smoothed average over recent reviews), so a 0.5-point single-week move is often the result of 30+ negative reviews posted that week — itself a signal that something material happened internally even if you can't see the underlying review content.

Related use cases

• Scrape Glassdoor salaries for compensation benchmarking
• Find Glassdoor interview questions by role
• Track Glassdoor rating changes over time
• The complete guide to scraping job boards
• All Thirdwatch use-case guides

Frequently asked questions

What does the actor return when scraping reviews?

Set scrapeType: "reviews" and the actor returns company-level review summaries with company_rating, pros (employee-reported positive themes as an array), and cons (negative themes), plus the company name and Glassdoor URL. Note this is summarised theme data rather than individual review text — Glassdoor's UI clusters common phrases. For deep individual-review analysis, you'd need a separate detail-page scrape.

How does this compare to AmbitionBox for Indian companies?

Glassdoor coverage outside US/UK is materially thinner. AmbitionBox has 5-10x more reviews per Indian company than Glassdoor for the same employer. For Indian employer-brand research specifically, use AmbitionBox; for global comparisons including US/UK companies use Glassdoor; for the full picture run both and merge by company name. Our AmbitionBox Scraper ships the seven-category breakdown that matches Glassdoor's review structure.

Can I track competitor culture over time?

Yes. Schedule weekly review snapshots and persist as Parquet. Compare company_rating across snapshots — a 0.2-point drop within 30 days is meaningful; 0.5+ points is rare and usually corresponds to leadership change, layoff, or PR crisis. Pros/cons text turnover is a weaker signal because Glassdoor's clustering algorithm changes mappings periodically; rely on company_rating drift as the primary trend metric.

What countries does the actor cover for reviews?

Seven Glassdoor country domains: US, UK, India, Canada, Australia, Germany, France. Pass the country code as the country input. Coverage is densest in US (1.4M+ companies indexed) and UK (300K+); India coverage exists but is materially thinner than AmbitionBox; smaller markets (DE, FR, AU, CA) have decent coverage for tier-1 companies and sparse data for SMBs.

How fresh is review data?

Each run pulls live from Glassdoor at request time. Reviews update as new employees post — typically 1-10 new reviews per week for active large employers. Company ratings move slowly (0.1 points per quarter for stable companies). For employer-brand monitoring, weekly snapshots are sufficient; for crisis monitoring during active reputation events, daily is justified.

Can I correlate Glassdoor reviews with company performance signals?

Yes. A Glassdoor company_rating drop of 0.3+ points often precedes earnings misses, leadership departures, or layoff announcements by 2-4 weeks for public companies. Pair Glassdoor with our LinkedIn Jobs Scraper to detect simultaneous hiring slowdowns; the combined signal is materially stronger than either alone.

Run the Glassdoor Scraper on Apify Store — pay-per-record, free to try, no credit card to test.