Job posting data as a hiring signal
A company posting three backend roles and a head of infrastructure is telling you something it has not announced yet. Hiring data is one of the most reliable growth signals available, and unlike funding announcements it appears continuously rather than in bursts. The difficulty is that postings are scattered across aggregators, applicant tracking systems, and company careers pages, all with different structures and heavy duplication.
- Starting at
- from £1,200
- Refresh
- Daily, or twice daily for fast-moving sectors
- Access
- public web
What comes out of it
- Job title, normalised seniority, and function
- Company, location, and remote policy
- Posted date, and closed date where detectable
- Salary range where disclosed
- Source URL and applicant tracking system in use
What makes Job boards and careers pages awkward
Deduplication is harder than it looks because the same role legitimately appears on the company site, two aggregators, and a recruiter listing, with different titles and sometimes different locations. Matching on title alone merges distinct roles at large employers. The workable approach combines normalised title, company, location, and posting date into a fuzzy key, then keeps the company-sourced version as canonical since it is the one that stays accurate.
Most of the value sits in the applicant tracking systems rather than the big aggregators. Greenhouse, Lever, Ashby and Workable serve structured JSON on predictable URLs per company, so a careers page built on one is far cleaner to read than any aggregator. Detecting which system a company uses and reading it directly gives better data than scraping a board that has already mangled it.
What the signal actually tells you
Volume by function is the clearest signal. A company that goes from two engineering roles to eleven in a quarter is scaling, and usually did so after funding that may not be public yet. A company posting its first compliance or finance leadership role is professionalising. A sudden stop in posting, or roles being pulled, points the other way and is often the earlier signal.
Location tells you about expansion before any press release does. First posting in a new country is a market entry, and it typically appears months ahead of an announcement.
Seniority mix matters too. Heavy junior hiring alongside no senior roles reads differently from a run of staff-level and management postings, and for anyone selling into engineering organisations that difference changes the pitch.
How it is collected
The pipeline starts by resolving each target company to its careers infrastructure, detecting which applicant tracking system it uses. Where a structured endpoint exists it is read directly, which gives clean titles, locations, and dates without parsing rendered HTML. Where it does not, the careers page is parsed with a per-pattern extractor.
Postings are then normalised: titles mapped to a controlled vocabulary of function and seniority, locations resolved to a consistent geography, and remote policy inferred from both the field and the text. Normalisation is what makes the data comparable across companies, and it is most of the work.
Closure detection runs by absence: a posting that disappears from a source is marked closed with a date, which is what lets you measure time-to-fill and spot roles being quietly pulled.
Scope and cost
Workflow automation starts from £1,200 for collecting vacancies from one agreed careers source into a defined output. A company watchlist spanning different careers systems, daily scheduling and Slack delivery is scoped separately.
Broad market coverage across an entire sector, with deduplication across sources, closure tracking, and trend analysis, is quoted separately against the agreed sources, delivery requirements and acceptance checks. The extra cost is mostly deduplication and the long tail of careers pages that do not use a standard system.
This category needs maintenance. Careers pages are redesigned regularly and applicant tracking systems change their public endpoints without notice. Plans start at £400 a month, and for a system anyone relies on operationally I would treat it as part of the running cost.
Where the line is
Publicly accessible pages, robots.txt and terms respected.
I work from publicly accessible listings and respect robots.txt and terms per source. Several large aggregators explicitly forbid automated collection, and for those I use their official APIs where offered or leave them out and say so, rather than quietly working around it. Company careers pages and the public endpoints of applicant tracking systems are generally fair game and are also better data. Postings sometimes name a hiring manager, which is personal data, so retention rules apply the same way as anywhere else.
MathersLabs is not affiliated with or endorsed by Job boards and careers pages. Work is scoped to publicly accessible data and to compliance with applicable terms and law, and feasibility is assessed per engagement before any fixed price is agreed.
Related work
SignalSauce
250 qualified leads a month, fully automated
Built an end-to-end intelligence pipeline that ingests from 5 curated news sources twice daily, scores every announcement against a custom algorithm, and enriches the ones that pass with 21 structured data points per company.
Questions people ask first
Is your source on this list?
Tell me what data you need and where it lives. I will tell you whether it is feasible, what it would take, and whether it is worth building before you commit to anything. Working hours overlap with US Eastern and Pacific.