Method
How it works
No scraping tricks, no private data, nothing bought from a third party. The whole method is boring on purpose, because that is what makes it repeatable.
Where the data comes from
Most technology companies run their hiring on one of a handful of applicant tracking systems, and each of those publishes an open feed of that company's live job postings so the company's own careers page can display them. Greenhouse, Ashby and Lever between them cover every company in this index.
Those feeds are public. Reading them is the same act as opening the company's careers page in a browser, and OrgTicker reads each one once a day, gently, at a fixed hour.
What that gives is a complete and honest picture of one thing: every role a company had listed on a given morning, with its title, department, location, employment type, and where the company chooses to publish it, the salary range.
What happens each morning
-
The index is read
Every company in the index is fetched and its full list of open roles is stored with a date stamp. Nothing is deleted, so yesterday stays exactly as it was.
-
Today is compared to yesterday
Roles are matched between the two days by a normalised title, which survives the small edits companies make to their own postings. What is left over is what genuinely appeared or disappeared.
-
Seven rules look for meaning
A change is only a signal if it says something. Each rule below has a threshold, and a change that does not clear it is recorded but never reported.
-
Your report is cut from that
Filtered to your vertical, your countries and the functions you fill, with a hard cap of three roles per company so a single large employer cannot fill your page on its own.
The seven signals
| Signal | What it means |
|---|---|
| First in a function | The company posted a role in a department it has never hired into before. The strongest signal there is, because the budget existed before the job did. |
| First in a country | A company's first ever posting in a new country. Market entry shows up in hiring before it shows up in a press release. |
| Scaling | Open roles up sharply over two weeks. Usually follows funding, a large contract, or a new product line. |
| Slowing | Open roles down sharply over two weeks. A drop of that size is a decision, not attrition. |
| Day over day | A meaningful one day move in either direction, reported with the function carrying it. |
| Seat reopened | The same role posted again shortly after it came down. The person did not stay, or never started. |
| Range moved | A role reposted with a different salary band. Compensation shifts before any survey reports them. |
What this does not do
- It does not tell you who works somewhere, who left, or how to reach anybody. There is no personal data in the system at all.
- It does not read anything behind a login, and it does not touch a company's site outside its published feed.
- It does not guess. Where a company does not publish a salary, the report shows no salary rather than an estimate.
- It does not cover every company in the world. It covers 168 and says which ones, which is the honest version of the same thing.
See it on your own patch
Tell me the vertical, the countries and the functions and I will build next Monday's report to exactly that.