LinkUp and Fokals work with hiring data the same way in one respect: both take job postings where the employer publishes them, not from the job boards that aggregate them. They differ in what each is built for. LinkUp's product page says its flagship dataset holds global job postings indexed daily since 2007, and it lists full-text job descriptions among the files. Fokals delivers every posting already labelled by job function, seniority and ten role flags, with daily Hiring Activity for each company, on the same company index as its technographic, intent and announcement data.
This page sets out what LinkUp offers and what it is built for, lines the two records up field by field, lists four other sources of job market data, and says which fits which use. Statements about other companies were checked against their own pages on 4 October 2026 and are linked. Fokals is delivered direct, by REST API and as bulk files, which you load into your warehouse with its own loader.
What LinkUp offers
LinkUp sells job market data collected from employer websites. Its products page describes five products.
- RAW. The flagship dataset: active and historical jobs back to 2007, as job records with full-text descriptions, structured fields, aggregated job counts and company reference files.
- Custom Feeds. Job listings delivered daily to the client's parameters. The feeds page names CSV, XML and JSON files.
- Market Reports. Reports, packaged or made to order, at the level of the economy, a region, an industry or a company, on a schedule the client chooses.
- Compass. A visual tool for querying and charting the data.
- JMD-Cloud. An archive of global labour demand that the page describes as unrevised, for point-in-time analysis.
A job record, as the RAW page lists it, carries the title, company and location, the dates it was posted, updated, removed and last checked, a SOC and O*NET occupation code, compensation and the description. Each company record carries a ticker and, the data page says, SEDOL, CUSIP, ISIN, LEI, PermID and NAICS code. LinkUp's data dictionary, last updated in June 2026, also documents point-in-time reference files: company and ticker rows with a start and an end date, which it says can be joined to job records to follow corporate changes such as mergers and acquisitions, name changes and web address changes.
The RAW page lists delivery through Snowflake, AWS, Azure, Google Cloud, FTP, Databricks and Google BigQuery. LinkUp's about page says it was acquired by GlobalData PLC in 2024.
What LinkUp is built for
Each line restates the pages linked above.
- A long record. Its product page says job postings have been indexed daily since 2007, and a knowledge base article of October 2022 dates job descriptions from 2014.
- Description text. Full-text job descriptions are among the files of RAW.
- Occupation codes. Each job record carries a SOC and O*NET occupation code.
- Point-in-time reference files. Company and ticker reference files with start and end dates, and JMD-Cloud, an archive its products page describes as unrevised.
- Delivery through cloud platforms. The RAW page lists Snowflake, AWS, Azure, Google Cloud, FTP, Databricks and Google BigQuery.
- Reports and a visual tool. Market Reports on a schedule the client chooses, and Compass for querying and charting the data.
Why a team looks at alternatives
The reasons are about fit, not about the data being right or wrong.
- Scope. The question is wider than postings: the same companies' technology, announcements and buying intent, on one company key.
- A different kind of data. The question is about the workforce a company already employs. Profile-based workforce data is built for that, and the guide to Revelio Labs alternatives covers it.
- Delivery. The pipeline wants a REST feed it can resume from a cursor, with bulk files for the first load.
- Labels. The model wants postings already labelled by job function, seniority and role, or it wants the occupation codes it already uses.
- Licence fit. The data is going into a product or on to clients, and the agreement has to name that use.
LinkUp and Fokals side by side
Each cell restates what the vendor says publicly.
| Aspect | LinkUp | Fokals |
|---|---|---|
| Scope | Job market data: job records, feeds, reports and a visual tool | Hiring beside technographic, intent and announcement data and weekly Market Series, on one company index |
| Source of postings | Employer websites | First-party: the roles each company publishes on its own careers pages |
| Job record | Title, company and location, posting dates, a SOC and O*NET occupation code, compensation and the full-text description | Every role labelled by job function, seniority and ten role flags, with location, work mode, advertised pay on one annual US-dollar scale and the tools named |
| Aggregates | Aggregated job counts among the files of RAW | Hiring Activity: daily open, new and closed postings per company; weekly Sales Team Metrics; Sales Pay Benchmarks by role and country |
| Company identifiers | Ticker, SEDOL, CUSIP, ISIN, LEI, PermID, NAICS code | Ticker, MIC, ISIN, LEI and share-class FIGI, and the SEC CIK where there is one |
| Point-in-time design | Company and ticker reference files with start and end dates; JMD-Cloud, described as an unrevised archive for point-in-time analysis | Every observation dated; daily and weekly datasets written once after the period closes and never revised |
| Delivery | Snowflake, AWS, Azure, Google Cloud, FTP, Databricks and Google BigQuery | REST API of 25 endpoints; bulk files as JSON, JSON Lines or CSV, with a manifest, delivered direct |
| Licensing | Its demo page says a demo leaves you with clear next steps for trial, legal and compliance | Written agreement for internal use, embedding in a product or redistribution |
How the two records line up
A team that evaluates both, or moves a model from one to the other, has to map the dates first. LinkUp's fields below are those of its data dictionary, given in words. The Fokals fields are those of Job Postings in the hiring dataset.
| Question | LinkUp RAW job record | Fokals Job Postings |
|---|---|---|
| When was it first seen? | The created date | The first-seen date, with a baseline flag on a posting already open when the company was first observed |
| When was it last seen? | The last-checked date | The last-seen date |
| When did it come down? | The delete date | The closing date |
| What identifies it? | A hash, with a base hash for the posting on the employer's site | A stable posting ID |
| Which company? | A company ID, joined to the reference files | The Fokals company ID, with the listing identifiers on the same record |
| What kind of job? | An O*NET occupation code | Job function and seniority labels, and ten role flags |
Three differences change a count.
Locations. LinkUp's dictionary says its dataset splits out a job for each location listed on a posting, so one base hash can have several hash values. Fokals delivers one record per posting, with every location listed on it. A posting for five cities is five records in one file and one in the other.
Baselines. When Fokals first observes a company, every posting already open sets the baseline and carries a flag, so the count of new postings holds true openings. If you rebuild a daily series from Job Postings, apply the same filter.
select
company_id,
cast(first_seen_at as date) as day,
count(*) as opened
from job_postings
where found_on_first_read = false
group by company_id, cast(first_seen_at as date)
order by company_id, day;That is the definition behind new postings in Hiring Activity, which is delivered already built for each company and closed day, beside open and closed postings and the open postings by function, seniority, country and work mode.
Closures. Each vendor defines when a posting has gone. Fokals confirms a closure before it is written, so a posting that drops out briefly and returns stays open in the record. LinkUp's dictionary defines the delete date by the latest read of the site that did not find the posting.
For these reasons, calibrate before you combine the two. Run both over the same weeks, measure the ratio company by company, and carry that ratio into any model that reads both. Acme Robotics, an illustrative company, could show 60 open postings in one file and 75 in the other on the same day and both be right. The wider method is in using hiring data in equity research, and the case for working at the source is in why we read job postings from the company's own board.
What Fokals builds on the posting record
Three datasets are delivered already built from Job Postings, so a team starts from measures and not from raw postings. Hiring Activity gives daily open, new and closed postings for each company, with the open postings by function, seniority, country and work mode. Sales Team Metrics gives the weekly shape of each company's sales organisation, including the first sales posting in a new country and the first enterprise-segment role. Sales Pay Benchmarks gives pay quartiles for sales roles by role and country, from advertised pay on one annual US-dollar scale.
Every one of them shares the company ID with the rest of the index. A rise in postings can be read beside Technology Changes, Intent Scores and Company News for the same company, and the weekly Market Series set one company's hiring against its industry, country and size band on same-store cohorts.
Other sources of job market data
None of these was tested for this guide: each line says only what the vendor states on its own site.
- Lightcast. Lightcast describes labour market data built from job postings, professional profiles and government statistics, delivered by API, cloud storage, data shares and SFTP.
- Revelio Labs. Revelio Labs offers COSMOS, a deduplicated postings dataset that it says is sourced from employer websites, major job boards and staffing firm boards, beside workforce data built from professional profiles.
- Coresignal. Coresignal describes job postings from the public web as deduplicated active and historical listings, by API or as datasets in Parquet, JSON Lines or CSV, beside its company and employee records. The guide to Coresignal alternatives covers it.
- PredictLeads. PredictLeads describes a job openings dataset sourced from company websites, career pages and applicant tracking systems, delivered by API, flat files and webhooks. The guide to PredictLeads alternatives covers it.
Which source for which job
| If you need | Consider |
|---|---|
| A back-test over many years of postings | LinkUp, whose job records start in 2007 |
| The text of job descriptions | LinkUp or PredictLeads, which both list descriptions among their fields |
| Headcount and the people already employed | Revelio Labs or another profile-based source |
| Labour market data built from postings, professional profiles and government statistics | Lightcast |
| Labelled postings joined to technographic, intent and announcement data on one company ID | Fokals |
| Sales hiring metrics and pay benchmarks by role and country | Fokals: Sales Team Metrics and Sales Pay Benchmarks |
| Daily monitoring of a coverage list of listed companies | LinkUp or Fokals: test both on your own list |
Frequently asked questions
How far back does LinkUp job data go?
LinkUp's product page says its flagship dataset holds job postings indexed daily since 2007. An article in its knowledge base, last updated in October 2022, says LinkUp began indexing job descriptions in 2014 and gives approximate coverage for the years after. Its market reports page lists history from 2012. Ask LinkUp for the current figures for the files you plan to license.
Can Fokals hiring data be used point-in-time?
Yes. Each posting carries its first-seen and last-seen dates, and Hiring Activity is written once after each day closes and is never revised, so a query as of any day the record covers returns what was known that day. A first observation sets a baseline and is never counted as a new posting. Labels are produced under named, frozen versions, with at least 90 days' notice of a breaking change, which keeps the record stable for research and for model features.
What is the difference between job postings from employer sites and from job boards?
A posting on an employer's own site is published once, by the company that is hiring, with its own dates. A job board or aggregator republishes postings from many employers, so the same job can appear several times. LinkUp and Fokals both work from employer sites. Vendors that draw on several sources, such as Revelio Labs and Coresignal, describe their postings datasets as deduplicated.
Does LinkUp map job postings to stock tickers?
Yes, according to its own pages. LinkUp's data page says each company record contains a ticker and other identifiers, including SEDOL, CUSIP, ISIN, LEI and PermID, and its data dictionary describes a ticker reference file with start and end dates for point-in-time joins. Fokals records carry ticker, MIC, ISIN, LEI and share-class FIGI for a listed company, and a brand or subsidiary carries the identifiers of its listed parent.
What labels does Fokals put on a job posting?
Every posting is labelled by job function, seniority and ten role flags, and carries its location, work mode, advertised pay on one annual US-dollar scale and the tools it names. The labels are produced under a named, frozen version, so a series built on them stays comparable over time. Hiring Activity then counts open postings by function, seniority, country and work mode for each company and day.
How do you load Fokals hiring data into Snowflake or Databricks?
Fokals is delivered direct, by REST API and as bulk files in JSON, JSON Lines or CSV, which you load with the platform's own loader. A first load comes from a bulk export with its manifest, and the feed of postings, which runs oldest first from a time you set, keeps the warehouse current from the last cursor. LinkUp's RAW page lists delivery through Snowflake, AWS, Azure, Google Cloud, FTP, Databricks and Google BigQuery.
The queries and code on this page are examples to adapt. Test them in your own environment before you rely on them.
What this page says about the products it names was checked against their public documentation on 4 October 2026. Product and company names are trademarks of their owners. Fokals is not affiliated with them or endorsed by them.