WARN Act Layoff & Plant-Closure Signal Feed — Layoff Data Scraper & API

A WARN Act notice scraper and layoff data API that pulls official, government-filed WARN Act (Worker Adjustment and Retraining Notification) notices directly from state labor agency sources and normalizes them into one clean feed: company, location, headcount affected, notice date, effective date, and closure type. No news aggregation, no scraping of secondary trackers — this reads the primary source filings state agencies are legally required to publish.

Use it to build a mass layoff tracker, a plant closure dataset, a sales/recruiting trigger-event feed, or a distress-signal input for credit/equity research — all from the same normalized WARN Act data feed.

Why this is worth charging for

Mass-layoff and plant-closure notices are one of the cleanest leading indicators of corporate distress available anywhere, and they are legally mandated public record — but there is no unified API for them. Every state publishes its own WARN notices in its own spreadsheet format, on its own schedule, with its own column names, and reformats them without warning. Aggregating this by hand across even a handful of states is a standing, recurring chore. That gap is the product.

Buyers who have historically paid for exactly this kind of signal:

• Recruiters and staffing agencies — a fresh, large layoff is a sourcing list.
• Asset recovery / equipment liquidators and commercial real estate brokers — a facility closure means fixtures, equipment, and space are about to hit the market.
• Distressed-debt and short-side investors — mass layoffs are a real-time input into credit and equity risk models, well ahead of earnings.
• Journalists and local economic-development offices — first-to-know on regional job losses.
• Sales teams selling into HR / outplacement / benefits — a layoff event is a textbook trigger-based lead.

Output

Each dataset item is one normalized layoff/closure notice:

{
  "state": "CA",
  "company": "Acme Manufacturing Inc",
  "noticeDate": "2026-01-05",
  "effectiveDate": "2026-03-10",
  "employeesAffected": 1250,
  "layoffType": "Layoff",
  "county": "Los Angeles",
  "city": "Los Angeles",
  "address": "100 Main St",
  "zip": null,
  "naicsCode": null,
  "region": "LA/Ventura",
  "sourceUrl": "https://edd.ca.gov/siteassets/files/jobs_and_training/warn/warn_report1.xlsx",
  "sourceSheet": "WARN Report",
  "headerDetectionConfidence": "high"
}

Input

Field	Description
states	Which state sources to pull (CA, TX at launch — see Roadmap).
sinceDate	Only return notices on/after this date (matched against effective date, falling back to notice date). Leave blank for full available history.
maxItems	Stop after this many normalized notices. Leave blank for no cap.
includeRawRow	Attach the untouched original spreadsheet row for audit/debugging.

Recommended pricing (Pay-Per-Event)

This actor is built for Apify's Pay-Per-Event model:

• state-source-checked — $0.05, charged once per state per run, regardless of how many new notices were found. A "quiet" check is still a monitored result; the user is paying to know nothing changed, not only for rows.
• layoff-notice — $0.10 per normalized notice pushed to the dataset. This is the primary event.
• apify-actor-start (synthetic, Apify-managed) — covers the first 5 seconds of compute per run.

A typical scheduled daily run across 2 states with a handful of new filings costs a buyer $0.20–$2.00; a full historical backfill across both states (several thousand historical rows) is a one-time $150–$400 pull — priced like the alternative-data product it is, not like a generic scraper.

Why the parser doesn't hard-code column positions

State agencies rename, reorder, and add spreadsheet columns without notice — the actor detects the header row and maps columns by keyword match instead of fixed index (see src/parser.js). If a sheet's header can't be confidently identified, that sheet is skipped rather than silently emitting garbage rows, and every record is tagged with headerDetectionConfidence so consumers can filter on it. Run npm test for the offline self-test that exercises this logic against synthetic CA- and TX-shaped workbooks.

Roadmap

Built to add states without changing the parser: add an entry to src/sources.js and, if that state exposes downloadable WARN data, it should work with no further changes. New York, Florida, and Ohio are natural next additions once their file formats are confirmed stable.

FAQ

Is there an API for WARN Act layoff notices? Yes — this actor is that API. Run it on a schedule and every new WARN notice across your selected states lands in a structured dataset you can pull via the Apify API, webhook, or direct export (JSON/CSV/Excel).

Where does the layoff data come from? Directly from each state's official labor agency WARN filing page (e.g. California's EDD, Texas Workforce Commission) — the same primary-source spreadsheets employers are legally required to file, not a news aggregator or third-party tracker.

Can I filter by date or company size? Yes. Use the sinceDate input to pull only new notices, and filter the output dataset by employeesAffected after the run — every record includes the full headcount, so you can slice by mass-layoff threshold yourself.

How much does it cost to run? Pay-per-event pricing: $0.05 per state source checked, $0.10 per layoff notice returned. A daily monitoring run across a couple of states typically costs well under $1; a full historical backfill is a one-time cost scaled to how many notices exist.

What's the difference between this and a layoff news tracker? News trackers report layoffs after they're publicly announced or leaked. WARN notices are a legal filing requirement with fixed advance-notice windows, so this feed is often the earliest structured, verifiable signal available — before it's a headline.