Skip to content

About

Why WorkforceSignal exists

We started this because the only way to follow tech layoffs was doomscrolling. So we built a tracker with a source on every number, and kept going from there.

The story

WorkforceSignal started from a job search. Watching hiring freezes spread, entry-level postings vanish, and hundreds of applicants pile onto every opening made one thing clear: something structural was happening, a collision of COVID-era over-hiring corrections, the end of cheap money, and the arrival of AI that actually does work. But the information about it was scattered across headlines, earnings calls and anecdotes.

Inspired by trackers like layoffs.fyi, this project goes a step further: not just counting the layoffs, but explaining them, separating AI reality from AI narrative, and turning the analysis into practical tools, a resilience calculator, skills roadmaps, and honest forecasts, for the people living through the shift.

Since September 2026 the site also tracks the other side of the ledger. A layoff tracker on its own is a mood, not a market: the same year that passed 2025's layoff total also saw software postings climb for fifteen months and three quarters of tech leaders plan to hire. The Opportunities page and the monthly briefing exist so that the upside is reported with the same sourcing discipline as the cuts, never as spin.

How the layoff tracker is curated

  • Sources. Company announcements, SEC filings, and credible reporting (Reuters, Bloomberg, CNBC, FT and peers). Every event names its source.
  • Scope. Major, named events, currently 126of them, rather than exhaustive coverage. Industry-wide totals (like layoffs.fyi's) run higher because they include hundreds of smaller companies.
  • Counts.Figures reflect what was reported at announcement time. Companies sometimes revise; where ranges were announced (e.g. "4,000–6,000"), we record a midpoint and say so in the note.
  • AI attribution. The AI flag is applied conservatively, only when the company itself or credible reporting tied the cut to AI (substitution, reallocation toward AI, or AI-driven efficiency). Narrative-only claims are discussed in our analysis, not silently counted.
  • Updates. Dataset last updated September 2026. The site is built so new events are added in one place and every chart and stat updates automatically.

The calculator, honestly

The Career Resilience Calculator is a transparent weighted heuristic. Its role-exposure estimates are informed by published task-automation research and three years of layoff patterns; its weights are documented on the page. It runs entirely in your browser, we never see your answers. It is a thinking tool, not a prophecy: use it to find your weakest factor and fix it.

How the numbers work

Two measures, never mixed. Our curated totals count only major, individually sourced events, so they run lower than industry-wide aggregates like Layoffs.fyi, which include hundreds of smaller companies. That gap is deliberate. When we cite an industry-wide figure, it always carries its source and a retrieved-on date; every chart, counter and index on this site is computed from the curated set alone.

Labor-market figures. Hiring intent, postings indexes, wage premiums and job-creation figures on the Opportunities page and in the monthly briefing are external measurements (Indeed Hiring Lab, Robert Half, the Bureau of Labor Statistics, PwC, LinkedIn and peers). Each carries its source and a date, is never averaged with our curated totals, and is refreshed monthly.

Undisclosed headcounts.Some companies confirm a cut without a number. Those events appear in the tracker with their percentage where known, but contribute nothing to any people-affected total, because we don't make up numbers. Wherever a total could be affected, we disclose how many events were excluded.

The Market Stress Index.Score = 50 × (mean monthly people affected over the last three complete months) ÷ (median monthly people affected over the twelve complete months before that), clamped to 4–97. The current month is excluded because curation backfills over weeks; a median baseline stops one mega-event from anchoring the denominator for a year; undisclosed-headcount events are excluded from both windows and their count is shown beside the dial. The implementation is open in the site's source (src/lib/stats.ts).

AI-attribution confidence. Events are graded: Confirmed means the company explicitly connected the cut to AI; Likely means multiple credible independent signals; Possible and Unsubstantiated are recorded in the data but never counted in any AI statistic. Headline AI-share figures count Confirmed and Likely only, and every AI share states whether it is measured by people affected or by events.

Independence

The dataset is the product, and it is not for sale. Three things this site will never do: accept payment for placement in, alteration of, or removal from the dataset; let any commercial relationship influence AI attribution or an event's framing; or present sponsored content as editorial.

What it may do, always with visible disclosure: carry clearly marked sponsored slots in the newsletter, use marked affiliate links outside the tracker and calculator, and offer paid access to data services. If any of that ever begins, it will be disclosed here first. Where we forecast, we state confidence levels and the evidence that would prove us wrong.

Corrections. When we get a figure wrong, the correction is published, dated and left visible, and the entry notes what changed and why. No correction has been required since launch; when one is, it will appear right here. To report an error, use the contact below and include a source.

Contact & contributions

Spotted an error, have a correction, or want to flag an event we missed? Email info@workforcesignal.com. Corrections are published, not buried.