How to Read a WARN Notice Without Misreading the Numbers

WARN data looks deceptively simple: company, place, date, and affected workers. Every field can carry qualifications. A notice may cover several worksites, list a range of separation dates, revise an earlier filing, or report positions rather than unique people. Reading the record responsibly means understanding what a number represents—and what it does not.

Start with the source, not the headline

A state spreadsheet or agency listing is often an index to the underlying notice. The index is useful for discovery, but the original filing usually contains more context: legal entity names, worksite addresses, job titles, union contacts, whether bumping rights exist, and whether the action is expected to be permanent.

Our tables retain an official-source link for this reason. Before citing a filing, open the state source and confirm that the indexed row still matches it. States revise files, replace spreadsheets, and publish amendments on different schedules.

The U.S. Department of Labor’s WARN compliance assistance page is the best federal starting point. It links the statute, regulations, worker and employer guides, and an interactive adviser. The Department notes that explanatory guides are not binding legal interpretations; courts enforce the federal Act.

Notice date is not always publication date

Three dates can be confused:

• Notice date: the date on the employer’s filing or the date the agency records as notice.
• Received or posted date: when the agency received or published the document.
• Effective date: when a closure, layoff, or first separation is expected to occur.

Some state registers publish only one or two of these. Others use a posting date in the main table and place the signed notice date inside a PDF. An amendment can have its own date while preserving the original effective schedule.

WARN Act Tracker calculates notice_period_days only when both a notice date and an effective date are present. That interval is a descriptive data point. It is not a verdict, because receipt dates, phased layoffs, exceptions, and amended notices can change the legal analysis.

“Affected employees” needs a denominator and a definition

An affected-worker count may refer to one worksite, one phase, one filing, or a revised total. It may include positions that are vacant by the effective date. A multi-location document can be represented as one filing by one state and several rows by another.

Avoid adding numbers until you check for:

• duplicate rows created by amendments;
• the same workers appearing in multiple phases;
• separate worksite rows that share a filing total;
• temporary layoffs later made permanent; and
• notices that were rescinded.

This is also why “workers affected” should not automatically be described as “people laid off.” A public notice reports an employer’s expected action at a point in time. Outcomes can change.

Worksite geography matters

The city in a state register may be the physical worksite, a corporate mailing address, or a regional label. Remote and traveling workers can raise additional questions about the relevant single site of employment. A company’s headquarters state is not necessarily the state where the affected work occurred.

For local reporting, confirm the street address in the original notice. For company-level analysis, group carefully across legal entities and name variations. “Acme, Inc.,” “Acme Holdings LLC,” and a consumer-facing Acme brand may or may not describe the same employer for legal purposes.

Layoff type is a classification, not a complete legal conclusion

State sources use inconsistent terms: closure, permanent layoff, reduction in force, relocation, furlough, or simply WARN. A normalized dataset has to map that language into a smaller set of categories. Our four public labels—plant closure, mass layoff, temporary layoff, and unknown—are intended for navigation.

The label should not be used to decide whether a statutory definition has been met. For example, duration matters for some temporary layoffs, and separate employment losses can sometimes be aggregated across a period. The federal regulations at 20 CFR Part 639 contain the controlling definitions and rules.

Amendments are information, not noise

An amendment may change the date, headcount, location, event type, or stated reason. Keeping only the latest total makes a clean chart but can erase the history of what workers and officials were told. Keeping every version without linking them can double-count the same event.

For serious research, treat a filing as a versioned record. Note the original date, amendment date, fields changed, and whether the agency identifies the filing as revised or rescinded. When an agency republishes a complete workbook, compare deterministic record identifiers and source values before overwriting prior data.

A citation-ready verification checklist

Before publishing a claim based on WARN data, answer six questions:

1. Is the source an official state agency or the original employer notice?
2. Does the legal employer name match the claim?
3. Is the date a notice date, received date, posting date, or effective date?
4. Does the headcount apply to this row alone?
5. Is there an amendment or rescission?
6. Are you describing a reported plan rather than a confirmed outcome?

Good WARN reporting is less about finding the largest number and more about preserving the qualifiers attached to it. A transparent dataset should make those checks easier, not hide them.