The market for job postings data is served by providers that deliver very different records under one name. This comparison is for an analyst at a fund or a data lead on a revenue team choosing among them. It sets six providers side by side on what their own documentation states: where the postings come from, what a record carries, and how a posting reaches a company and a listed security. After it you can open any sample and say why its counts differ from the next one's.
Each provider is described only from its own pages, read on 7 October 2026. Where those pages are silent the cell says so, a statement about the pages read and nothing more. No provider's figures for volume, history or price are repeated here. Fokals is one of the six, described from its public data dictionary, methodology and sourcing statement.
How the market for job postings data divides among providers
Providers of job postings data divide along three lines. The first is the source: some read employers' own sites, others combine company sites, job boards and aggregators and merge the copies. The second is the record: its dates, labels and pay. The third is the link from a posting to a company and to a listed security.
Each line changes what a count means. The source decides how many records one role produces, the dates decide what a series of new postings measures, and the link decides whether the data joins to a security master or a CRM on a key. The post on reading job postings at the source treats the first line in depth. The rows below rest on these pages.
- Lightcast. Its job postings methodology, company classification methodology, data page, API overview and the Postings, Company and Lightcast Companies tables of its data shares.
- Revelio Labs. Its data page and the datasets page of its data dictionary.
- LinkUp. Its RAW page, data page and data dictionary.
- Coresignal. Its job postings page and the dictionaries of its multi-source jobs data and company data.
- People Data Labs. Its job posting schema, data overview, data sources, daily deliveries, API quickstart and company schema.
- Fokals. The three documents linked above.
For one provider on its own, see the pages on LinkUp, Revelio Labs, Coresignal and People Data Labs alternatives.
Which job postings data providers collect listings directly from company career pages?
Three of the six document the employer's own site as the source of every posting: LinkUp names company hiring websites and People Data Labs company career pages, and Fokals collects the roles each company publishes on its own careers pages. Lightcast, Revelio Labs and Coresignal document company websites as one source among several, with copies merged.
| Provider | Where postings come from | How copies are handled | Delivery |
|---|---|---|---|
| Lightcast | Publicly available online job boards and company websites | Within each source, then across sources on normalised title, company and location | API, cloud storage, data shares, SFTP and software |
| Revelio Labs | COSMOS: employer websites, major job boards and staffing firm boards | Unified and deduplicated, with an indicator for each source a posting was found in | API, data feed and Terminal |
| LinkUp | Company hiring websites, directly | One job for each location listed on a posting, under a shared base identifier | Snowflake, AWS, Azure, Google Cloud, FTP, Databricks and Google BigQuery, and daily custom feeds |
| Coresignal | Multi-source dataset: job boards, company websites and applicant tracking systems | Cross-source records merged into one listing | Jobs API, and datasets as Parquet or JSON Lines |
| People Data Labs | Company career pages, exclusively | A persistent posting ID and the address of the posting | Job Posting Search API and daily files in an S3 bucket, both marked as beta |
| Fokals | First-party: the roles each company publishes on its own careers pages | One record per posting, with a stable posting ID and the address of the public posting | REST API, and bulk files as JSON, JSON Lines or CSV |
People Data Labs' schema and overview pages state that its Job Posting Data is currently in beta and that field definitions may still change. A count built from several kinds of source rests on the rule that merges copies. Lightcast treats the first advertisement of a job as the original posting, and advertisements for it found elsewhere in the next 60 days as duplicates. A count read at the employer's site rests on what the employer published.
Job postings read directly from company career pages and job board APIs: Greenhouse, Lever and Ashby
A company career page can be a job board hosted by the employer's applicant tracking system, and Greenhouse, Lever and Ashby are such systems. A provider that reads postings directly at the source may read those boards. Coresignal names applicant tracking systems among its sources, and each Fokals posting names the system whose board it came from.
Greenhouse, Lever and Ashby each present their product as applicant tracking software. Greenhouse's Job Board API returns a company's offices, departments and published jobs as JSON, and its documentation says the data is publicly available, so GET requests need no authentication. Ashby's job postings API returns all currently published postings of an organisation. Lever's developer documentation lists postings among the resources of its API, whose requests must be authenticated. Ask a provider that reads career sites which systems' boards it reads. The Fokals data dictionary lists the systems a posting can name, and the entry on the applicant tracking system explains why the board is the source.
Which job data providers offer an API for postings sourced from company career sites?
Lightcast, Revelio Labs, Coresignal, People Data Labs and Fokals each document an API for job postings, and LinkUp documents delivery through cloud platforms, FTP and daily feeds. Of the three sourced wholly from company career sites, People Data Labs documents a search API it marks as beta, Fokals a REST API, and LinkUp RAW files and custom feeds.
The Fokals feed of postings runs oldest first from a time you set, so the last cursor is the bookmark for the next run. API vs bulk files for company data sets the two routes side by side.
What a job posting record carries: dates, labels and pay
The word posting covers records with different dates. Read the date columns first, because they decide what a series of new postings measures.
| Provider | Dates and status | Labels | Pay |
|---|---|---|---|
| Lightcast | The date first posted, and expiry when a revisit finds the advertisement gone or by age | Occupation, title and skills on its own taxonomies | Advertised salary, extracted and cleaned |
| Revelio Labs | A post date and a remove date, empty while the posting is live | Role at several levels of detail, predicted seniority, remote type | Salary in annualised US dollars |
| LinkUp | Created, last checked, last updated and delete dates, from its own reads of the employer's site | A SOC and O*NET occupation code, education, experience and employment type | The compensation amount or range on the listing |
| Coresignal | Posting date, expiry date, created and updated times, and a status of active, expired or deleted | Seniority, management level, department, functions and employment type | Minimum and maximum advertised salary, with currency and period |
| People Data Labs | First seen, last verified and deactivated dates | Class, role, sub-role and levels derived from the title, remote work policy, inferred skills | Minimum and maximum salary as listed, with currency and period |
| Fokals | The employer's posting date, first-seen and last-seen dates, a closing date and a baseline flag | Job function, seniority and ten role flags, with a named version on the record | Pay as stated, and on one annual US-dollar scale |
Three differences in these records change a count or a model.
- Posted or observed. LinkUp defines its created date as the first time the job was observed. People Data Labs describes its first-seen date as the date it believes the posting was posted. Fokals carries the employer's posting date and its own first-seen date side by side. For point-in-time work, use an observed date.
- How a posting ends. LinkUp dates a removal by the latest read of the site that did not find the posting. Lightcast expires an advertisement when a revisit finds it gone, or after no more than 120 days. Fokals writes a closing date once the close is confirmed.
- Text and labels. Lightcast, LinkUp, Revelio Labs, Coresignal and People Data Labs list the posting's description or full text among their fields, which suits a team that trains its own classifier. Fokals delivers each posting labelled by job function and seniority, with a named version on the record.
Which alternative data providers map job postings data to stock tickers?
All six document a stock ticker in their data. Lightcast, LinkUp and Revelio Labs document company tables or reference files with a ticker and other security identifiers. People Data Labs and Coresignal document one on the company record. Fokals carries ticker, MIC, ISIN, LEI and FIGI on the posting itself.
| Provider | Company key on a posting | Security identifiers documented | Parents and change over time |
|---|---|---|---|
| Lightcast | A company ID and name from its Companies Taxonomy, which holds Tradestyle, NAICS codes and staffing labels | Ticker symbol, stock exchange, CUSIP, CIK and FIGI in its Company table; ISIN in its Lightcast Companies table | Subsidiaries whose names include the parent are generally counted under the parent company |
| Revelio Labs | Its own company ID on every posting | Ticker, exchange name, SEDOL, ISIN, CUSIP, CIK and LEI in the Company Reference file | The ultimate parent's company ID and name |
| LinkUp | A company ID that joins the job record to reference files | Ticker, SEDOL, CUSIP, ISIN, LEI and PermID | Reference rows carry a start date and an end date |
| Coresignal | Company name, website, domain and source-specific company IDs | A stock ticker and exchange in its company dataset | Parent company ID, name and website in the company dataset |
| People Data Labs | A company ID described as a foreign key to its company dataset | A ticker and the MIC of its exchange on the company record, for public companies | Immediate and ultimate parent IDs, and the ultimate parent's ticker and exchange MIC |
| Fokals | One stable company ID on every record | Ticker, MIC, ISIN, LEI and share-class FIGI on the posting, when the company or its parent is listed | A brand or subsidiary carries the identifiers of its listed parent |
A ticker is where the mapping starts. It names a listing on one venue, so a join to holdings is safer on an ISIN or a FIGI, as ISIN vs LEI vs FIGI sets out, and a back-test needs the mapping as it stood on each date. LinkUp's reference rows carry start and end dates for that purpose. Fokals marks a delisted listing and keeps it, so a join from an old position still resolves. The guide to mapping alternative data to a security master gives the order of keys.
Job postings API and company ticker mapping: Lightcast, Revelio Labs and People Data Labs
Lightcast, Revelio Labs and People Data Labs each document a job postings API. Lightcast matches the company name to its Companies Taxonomy and lists a ticker in a Company table. Revelio Labs puts its company ID on the posting and a ticker in a reference file. People Data Labs joins the posting to a company record with a ticker.
Lightcast's API overview says postings are enriched with standardised company names and can be filtered by company, and its data shares document a company ID on each posting and a Company table with a ticker symbol. Whichever API you call, test the join on the companies you track before you trust a total by ticker.
The questions that separate the datasets
Put the same questions to every provider.
For an investor
- When could the posting first be seen? A back-test needs an observation date kept apart from the employer's date.
- Is the link to a security dated? A mapping that holds today's ticker alone assigns old postings to today's owner.
- Do closed periods change? Compare two deliveries of the same days. Fokals Hiring Activity, the daily open, new and closed postings per company, is written once after each day closes and is never revised.
- Does new coverage look like new hiring? People Data Labs leaves its first-seen date empty for a posting already live at its first crawl, and a Fokals posting already open when a company is first observed carries a baseline flag and is never counted as new.
For a sales team
- Can the data key on an account's website? Each of the six documents a company website or domain: Coresignal and People Data Labs on the job record, Lightcast, LinkUp and Revelio Labs in company reference data, and Fokals in Technology Stack, beside the company ID.
- What does the posting say about a purchase? Look in a sample for tools named and role labels. Fokals postings carry the tools named and flags such as team build and budget owner, which timing outreach with hiring signals puts to work.
A test on a sample, with a worked example
Ask each provider for a sample of the same companies over the same weeks, and expect the counts to disagree. The example is illustrative: Acme Robotics is an invented company and every figure is an assumption.
Assumptions:
- Acme Robotics has 40 postings open on its own careers site on the test day.
- 6 of the 40 list three locations each, and the other 34 list one.
- Each of the 40 is also carried by two job boards.
- 5 roles closed on the careers site last week are still shown by one job board.
- 3 roles are advertised on a job board and never on the careers site.
| Counting rule | Working | Open on the test day |
|---|---|---|
| One record per posting, read at the employer's site | 40 | 40 |
| One record per location listed, read at the employer's site | 34 + (6 × 3) | 52 |
| Every listing found across the three sources | 40 + (40 × 2) + 5 + 3 | 128 |
| Those listings merged into one record per job | 40 + 5 + 3 | 48 |
Three delivered files could report 40, 52 and 48 open postings for one company on one day, each correct under its own rule. The second is 52 / 40 = 1.3 times the first, and the third 48 / 40 = 1.2 times. Then run five steps on the samples.
- Fix the list and the window. The companies you track, with the identifiers you hold, over the same closed weeks from every provider.
- Count the same thing. Take the ratio of open postings between samples, company by company and day by day. A steady ratio is a counting rule, as in the example. A ratio that moves is coverage or the dating of an ending.
- Trace records to the source. Open the address on twenty records from each sample, and note whether it leads to the employer's site or to a copy.
- Join on a key. Join each sample to your security master or CRM on the identifier its provider documents, and count the companies matched and left over.
- Take a second delivery. Compare the rows for the days both deliveries contain: a row that differs was revised.
Fokals sends sample data on request for the companies you name, with the data dictionary and the methodology. The hiring dataset page lists what each record delivers, and using hiring data in equity research takes the daily counts into a research design.
Frequently asked questions
Does Lightcast job postings data carry company identifiers or a ticker for a public company?
Lightcast's methodology says the company named in a job posting is normalised and matched to its Companies Taxonomy, where each company has metadata that includes Tradestyle, NAICS codes and staffing labels. The documentation of its data shares, read on 7 October 2026, puts a company ID from that taxonomy on each posting and lists a ticker symbol, stock exchange, CUSIP, CIK and FIGI in a Company table.
Does Revelio Labs job postings data map a company to a ticker?
Yes, through a reference file. Revelio Labs' data dictionary puts its own company ID on every job posting, with the ID and name of the ultimate parent. The Company Reference file, which it says is included with each delivery, lists a ticker and an exchange name beside SEDOL, ISIN, CUSIP, CIK and LEI.
Why do two providers show different counts of open postings for the same company?
Usually because they count different things. One file holds one record per posting on the employer's site, another one per location listed, a third one per job after merging copies from several sources. They also end a posting differently: when a read no longer finds it, when it reaches an age limit, or when a close is confirmed. Measure the ratio company by company over the same weeks before you combine two sources or move from one to the other.
How does Fokals tie a job posting to a listed security?
Every Fokals record carries one stable company ID. When the company or its parent is listed, the posting also carries the listing's ticker, MIC, ISIN, LEI and share-class FIGI, so a brand's postings can be added to those of its listed parent or kept apart. Postings join to a security master on ISIN, on FIGI, or on ticker with MIC, with no matching on company names.
Can job postings data be used point-in-time in a back-test?
It can when each record carries the date it was first observed and closed periods stay as written. Look for three things in a provider's documentation: an observation date kept apart from the employer's posting date, a dated mapping from company to security, and a statement on whether past periods are revised. Fokals Hiring Activity is written once after each day closes and is never revised, and a posting already open when a company is first observed carries a baseline flag.
What this page says about the products it names was checked against their public documentation on 7 October 2026. Product and company names are trademarks of their owners. Fokals is not affiliated with them or endorsed by them.