The number that finally changed how I read this market wasn't in a headline. It was 3 — the number of separate employers in King County that filed layoff notices the same week everyone I know was forwarding each other stories about Amazon layoffs. None of the three were Amazon. One was a logistics company in Kent I'd never heard of. One was a health system. One was a software company in Bellevue that had raised money eighteen months earlier.
I spent about three weeks refreshing a corporate count and learned almost nothing about my own odds. Then I opened a spreadsheet the state of Washington publishes for free and learned more in forty minutes.
The verdict, in one sentence: headline layoff counts from any single employer are close to useless for predicting your regional job market, and three free public data sources — state WARN filings, metro-level job posting trends, and downtown worker foot traffic — will tell you more about the next six months than every news alert you have set.
No affiliate links in this piece. Everything I name is free and public. I'm not selling you a dashboard.
What most people do
Here's the pattern I fell into, and I've watched maybe a dozen people I know fall into the same one.
A round of Amazon layoffs gets confirmed. The confirmation comes with a statement about focus and difficult decisions and continued investment in the areas that matter. Sometimes it comes with a number. Increasingly, as of writing, it doesn't — the company confirms the cuts and declines to say how many. Then everyone reads the statement three times looking for a tell, argues about whether "reorganization" means the same thing as "replacement," and closes the tab feeling worse and knowing nothing new.
Then comes the extrapolation, which is where the real damage happens. One large employer cuts, so the reader concludes the regional market is closed. Or the same reader notices the company is still hiring aggressively in one division and concludes it's fine. Both conclusions come from the same mistake: treating a single employer's internal capital allocation decision as a readout on a labor market of roughly two million people.
Amazon is enormous here and I'm not pretending otherwise. But Boeing, Microsoft, Costco, T-Mobile, Providence, Alaska Airlines, Nordstrom, Expedia, the University of Washington, and about four hundred companies you've never heard of also hire in this region, and they do not move in lockstep. In the same quarter I watched one big tech employer contract, a friend of mine with almost the same résumé took an offer at a healthcare payer in Renton for within a few thousand dollars of her old salary. The market wasn't closed. It had moved, and the headline had no way to tell her where.
The second thing most people do is track the wrong lag. A layoff announcement describes a decision that was made weeks or months earlier. By the time you read it, the hiring freeze that preceded it has already been in effect for a while, and the competitive pressure on job postings has already arrived. You are reacting to weather that has already passed through.
What the evidence suggests
There are three sources I'd actually build a picture from. Each one is genuinely good at something and each one will mislead you in a specific, predictable way.
WARN filings. Under federal law, larger employers have to give 60 days' notice of qualifying mass layoffs, and the notices go to the state. Washington's Employment Security Department publishes the list: employer, location, number of workers, effective date. It is the single most concrete layoff data that exists, and almost nobody outside HR reads it. Washington has also layered its own notification rules on top of the federal ones and the thresholds have shifted in recent years, so read the current rules on the state page rather than trusting my summary of them.
Where it misleads you: it undercounts, badly. Small cuts don't trigger it. Rolling cuts structured under the threshold don't trigger it. Severance paid in lieu of notice means people are gone long before the effective date on the filing. Remote workers get assigned to a site address that may have nothing to do with where they live. Treat WARN as a floor, never a total.
Metro-level job posting data. Indeed's Hiring Lab publishes posting trends by metro and by occupational category, free, no login. What matters isn't the absolute level, it's the shape: software postings can be down while healthcare, skilled trades, and logistics postings hold flat or climb. That divergence is the actual story of this region and it never fits in a headline.
Where it misleads you: postings are not jobs. Ghost postings, evergreen requisitions, and roles reposted after an internal candidate falls through all inflate the count. A rising line means demand is not collapsing. It does not mean there are that many real openings.
Downtown worker foot traffic and office occupancy. The Downtown Seattle Association publishes worker return metrics. If you own a business downtown, this is your leading indicator, not the tech news. Lunch revenue tracks badge-ins, not headcount announcements.
Where it misleads you: foot traffic conflates workers, tourists, and residents, and a company can shed thousands of roles without changing its badge-in count much if the cuts were remote or distributed across other sites.
| Source | Cost | Best at | Where it lies to you | Lag |
|---|---|---|---|---|
| State WARN filings | Free | Confirmed, dated, employer-specific cuts | Undercounts small and rolling layoffs | Filed ahead of the date, but decisions are older |
| Metro job posting data | Free | Which sectors are absorbing people | Ghost and evergreen postings inflate it | Roughly a month |
| Downtown foot traffic | Free | Consumer spending and small-business demand | Mixes workers with tourists and residents | Monthly, sometimes quarterly |
The part that matters for business owners: layoffs at large employers show up in local consumer spending later and more unevenly than people expect. A laid-off worker with several months of severance does not change their spending immediately. They change it around month four, when the search has gone longer than they planned. That's the delay to plan inventory and staffing around.
What I actually do
Once a month, on the first weekend, I spend about forty minutes on this. Not daily. Daily is how you end up with a nervous system tuned to press releases.
First, I pull the state WARN list and filter to the last 90 days in King, Snohomish, and Pierce counties. I count two things: total filings, and filings from employers that are not the one currently in the news. The second number is the one I care about. When it climbs, that's a broad contraction. When it's flat and one giant is doing all the cutting, that's a single company restructuring, and my search strategy shouldn't change much.
Second, I check posting trends for my occupational category against two adjacent ones — for me that meant backend engineering against data infrastructure and against IT operations at non-tech employers. When my category is down and an adjacent one is flat, that's a signal about which job titles to put on applications, and it is the most directly actionable thing on this list.
Third, I look at foot traffic once a quarter, not monthly. It moves too slowly to be worth more attention than that.
The thing I stopped doing: tracking any single company's headcount. I did it for months. It produced anxiety and no decisions. The information had no path to an action I could take.
I won't pretend this routine got me hired. I sent 419 applications and got 4 callbacks, and none of the four came from a clever read of a data set. What it did was stop me from making two expensive mistakes: sitting out a month because I thought the market was closed when it wasn't, and aiming exclusively at employers whose hiring had frozen a full quarter before the news said so.
Who this is for, and who it isn't
This is for you if you're job-hunting in the Puget Sound region and keep getting told, by people with a network they didn't have to build, to network more. It's for you if you own a business whose revenue depends on people coming downtown. It's for you if you're deciding whether to relocate, retrain, or wait.
This isn't for you if you already have an offer in hand — take the offer, close the spreadsheet. It also isn't for you if you're looking for a forecast. None of these sources predict anything. They describe what already happened with less delay and less spin than the alternative, which is a different and more modest thing.
Tonight, before you refresh one more headline: open your state's WARN list, filter to the last 90 days, and count how many filings are from employers nobody is writing about — that number, not the one in the news, is the market you're actually in.
