Job posting database: a buyer’s evaluation guide
A practical guide to testing coverage, record structure, deduplication, company matching, history, delivery, and interpretation with your own records and queries.
- written by
- Adewale ThomasMixRank
- published
- 19 November 2026
- reading time
- 11 min
- filed under
- Labor Market Intelligence
A job posting database gives you structured records of advertised roles for data products, research, and analysis. You can use those records to study hiring activity by employer, role, location, and period.
Start your evaluation with the work you need the data for. A labor-market research project may care most about historical coverage and consistent classifications. An HR technology product may care more about current statuses, structured fields, and API access. For example, a team comparing 25 employers over 12 months can build its test set around those exact companies and periods.
This guide shows you how to test a job posting database with your own records and queries. It covers record structure, deduplication, company matching, historical data, delivery, and interpretation.
What is a job posting database?
A job posting database is a structured collection of current and historical job advertisements. Its records can include role details, employer details, locations, requirements, posting statuses, and timestamps. MixRank’s job posting database includes 1.5B+ total postings since 2021 and 50+ fields.
The database is the input for products and research. If you want the analysis process itself, read our guide to job posting analysis.
Providers differ in coverage, field structure, duplicate handling, company matching, historical depth, and delivery. Those differences should shape what you test.
Job posting database versus job board
A job board and a job posting database can show the same advertisement to very different users, for very different work.
| Dimension | Job board | Job posting database |
|---|---|---|
| Primary user | Job seekers and recruiters | Data teams, researchers, and product teams |
| Main action | Discover and apply for roles | Query, analyze, enrich, and integrate records |
| Typical output | Advertisements and application pages | Structured records and dataset deliveries |
| Historical access | Usually centered on available opportunities | May include closed postings and historical changes |
| Company analysis | Generally organized around individual employers | Can connect postings to persistent company records |
| Common access | Website or candidate-facing application | API, bulk files, or warehouse delivery |
| Buyer concern | Candidate experience and application flow | Coverage, data quality, history, and integration |
This distinction matters because people searching for this term can be candidates or data buyers. The rest of this guide is about evaluating structured job posting data for use inside your organization.
What records and fields should buyers expect?
A provider’s schema tells you which fields the database can contain. Your evaluation should show how those fields hold up on the records you actually need.
Record identifiers
Each posting should have an identifier your systems can store and reference. Ask what happens to that identifier when a posting changes, closes, reappears, or has related advertisements.
Stable identifiers let you process updates without creating duplicates in your own database.
Role and classification fields
Common fields include job title, department, function, seniority, occupation, and employment type. Check whether the classifications are standardized and whether the provider defines them.
Then check them against the roles you care about. Two postings titled “Data Analyst” can differ a lot in responsibilities and seniority.
Location and workplace fields
Location fields may include city, region, country, and workplace type. Test how the database handles postings with several locations, remote roles, regional advertisements, and incomplete addresses.
If you plan to compare regions, you’ll need location handling that stays consistent.
Compensation, benefits, and requirements
A schema may support salary, benefits, skills, qualifications, experience, and responsibilities. Not every posting fills every field.
Count the populated and missing values on your test set. That tells you more about what you can analyze than reading the schema alone.
Status and timestamps
Look for fields that record when a posting was published, updated, and closed, plus its current status. Ask what each status means and which timestamp changes when the provider updates a record.
These fields decide whether you can reproduce active-posting counts and historical analyses.
Company context
Postings connected to company records let you analyze hiring by employer, industry, company size, and other firmographic attributes.
Ask which company identifier connects each posting to an employer record. Then test whether that connection holds across locations, subsidiaries, and name variations.
Metadata and change history
Metadata can include record IDs, posting URLs, and change-related fields. How much you need depends on your processing and audit requirements.
Ask for enough detail to trace a record through your own pipeline and investigate changes you didn’t expect.
How to evaluate a job posting database
Start with what your use case needs. The table below turns common use cases into database requirements.
| Use case | Requirements to prioritize |
|---|---|
| Labor-market research | Historical coverage, consistent classifications, and documented metric definitions |
| HR technology | Structured fields, current statuses, query support, and API access |
| Investment research | Company matching, historical changes, and employer coverage |
| Sales intelligence | Current postings, usable company identifiers, and relevant filters |
Use those requirements to weigh the criteria below.
1. Relevant coverage
List the employers, roles, locations, and periods your workflow needs. Ask each provider to show coverage against that list.
Break the results down so you can see strengths and gaps within your target group. One total count can hide big differences, such as strong coverage for large employers and thin coverage for smaller ones.
2. Deduplication
Repeated advertisements can inflate posting counts and skew employer comparisons. Ask how the provider represents related advertisements and deduplicated records.
The U.S. Bureau of Labor Statistics (BLS) notes that employers can advertise positions across numerous platforms and locations. Put a few of these in your test set, such as one role advertised on several sites or in several cities.
3. Expired-posting handling
Some online listings stay visible after the position is filled. Check how the database marks closed, expired, or removed postings.
Test whether the status fields and timestamps support your definition of an active posting.
4. Company matching
Evaluate company matches using employers with similar names, subsidiaries, multiple locations, and naming variations. For example, include two unrelated employers that share a common name.
A wrong company match carries through every later step of your analysis. Record both missed matches and incorrect matches during testing.
5. Field completeness and consistency
Check the fields you need on your own records. Count missing values and review formats for dates, locations, classifications, salary values, and identifiers.
Consistent records will reduce the cleanup your team needs to do.
6. Historical depth and update timing
Historical depth sets how far back your research can go. Update timing sets when new and changed records reach you.
Write down both requirements separately. Confirm that closed postings stay available for the full historical period you need.
7. Search and query support
Test the filters and identifiers you need to isolate your target records. Try combinations of employer, role, location, status, and date. For example, ask for open engineering roles at five named employers in Texas, posted in the last 90 days.
For API workflows, review pagination, batch behavior, error responses, and documentation. For analytical workflows, look at sample exports and how you’d query history.
8. Delivery and integration
Check the delivery method against the system that will use the data.
A job posting data API can support search and product workflows. Bulk files or direct warehouse delivery can support larger analytical workloads.
Your implementation estimate should include transforming responses, matching companies, processing updates, and monitoring failures.
Using bulk job postings data for hiring signals
Bulk job postings data can help you spot changes in hiring activity across a group of employers. For example, you might compare how many customer support roles 20 retailers advertise each month. To trust that comparison, you’ll need consistent company matching, deduplication, timestamps, and historical coverage.
Start by defining the group you’re measuring, such as a fixed set of roles and locations at each employer. Stable record identifiers, deduplication rules, company matches, timestamps, and posting statuses let you reproduce the count and explain why it changed.
Before you buy bulk data for this, decide:
- Counting unit: Will you count unique postings, posting versions, reactivated listings, or individual job openings?
- Employer scope: Check that postings resolve to consistent company records, including subsidiaries or regional entities where they matter.
- Time field: Pick the timestamp that drives the comparison, such as publication, first observation, update, or closure.
- Coverage: Test whether employer, role, and location coverage stays consistent across the full comparison period.
- Status handling: Decide how expired, removed, duplicated, and reposted records enter your calculations.
- Historical depth: Confirm the available history covers your baseline and comparison windows.
Job postings show advertised hiring activity. On their own, they don’t confirm completed hires or net workforce growth. Duplicate advertisements and listings that stay up after a role is filled can also skew simple counts. Read changes using consistent filters, documented counting rules, and relevant company context.
MixRank supports bulk hiring-signal workflows through flat-file exports and direct data-warehouse delivery, with job openings and hiring activity connected to the hiring company’s record. You can test these records against a fixed group of employers, roles, locations, and periods before you build production hiring-signal metrics.
How to test a provider with your own records
A structured test makes your provider comparison easier to defend.
Step 1: Define the test scope
Write down the companies, roles, locations, periods, and posting statuses your workflow depends on. Include common cases and hard ones.
Hard cases include similarly named employers, repeated advertisements, employers with few postings, roles posted in several locations, and records with known status changes.
Step 2: Fix the test records and queries
Use the same inputs with every provider on your shortlist. If providers need different query structures, note how you translated each request.
Run separate tests for record retrieval, company matching, historical coverage, and delivery.
Step 3: Inspect duplicates and statuses
Review repeated advertisements and see how each provider represents them. Check that your counting rules give the results you expect.
Check active, closed, expired, and historical records against your definitions.
Step 4: Score company matches and required fields
Define labels for correct matches, incorrect matches, and missing matches. Then check the fields your workflow needs on the correctly matched records.
Keep results by field and by test case. An overall score can hide a few failures that matter a lot, such as one large employer matched to the wrong company.
Step 5: Exercise the delivery method
Run representative API requests or process a sample delivery. Include pagination, empty results, invalid inputs, updates, and error responses where they apply.
Confirm that the delivered records match the documented schema and that your systems can process them reliably.
Step 6: Record implementation work
Estimate the work needed to transform fields, store identifiers, resolve companies, process updates, and monitor failures.
Add open questions to the evaluation record and give each one an owner. That keeps the buying decision tied to evidence.
How to interpret counts, statuses, and historical records
Clear definitions head off several common interpretation problems.
Choose a counting unit
Decide whether your analysis uses raw listings, deduplicated postings, active postings, new postings, closed postings, or hiring companies.
Each unit answers a different question. Label the unit you chose in charts, tables, and research notes.
Separate activity from workforce measures
Job postings describe advertised opportunities during a period. Employment statistics describe people working in jobs, while population-level skills data describes skills held by a wider group.
OECD guidance treats online job postings as flows of advertised opportunities and distinguishes them from workforce stocks.
Define active and historical records
Write down the statuses and timestamps that decide whether a posting counts as active in a given period.
Use the same definitions across providers and time periods wherever you can. Note any exceptions that could affect comparisons.
Track coverage and methodology changes
Coverage changes can move a time series. Read the provider’s documentation and look into unexplained shifts before you put them down to employer behavior.
Keep notes on classifications, filters, and processing decisions alongside the results.
Job posting database buyer’s checklist
| Criterion | Buyer question | Evidence to retain |
|---|---|---|
| Relevant coverage | Does the database cover our target population? | Segmented results from the test set |
| Record identifiers | Can we track records through changes and updates? | Identifier behavior on test records |
| Deduplication | How are repeated advertisements represented? | Duplicate-case examples and resulting counts |
| Expired postings | Can we apply our active-posting definition? | Status fields, timestamps, and known changes |
| Company matching | Are postings connected to the correct employer? | Correct, incorrect, and missing-match labels |
| Field completeness | Are required fields populated and consistent? | Field-level results by test case |
| Historical depth | Does the database cover our research period? | Historical records, including closed postings |
| Update timing | When do new and changed records arrive? | Observed updates and relevant timestamps |
| Query support | Can we isolate the records our workflow needs? | Representative requests and responses |
| Delivery | Can our systems process the supplied records? | API test results or sample-delivery review |
| Metric definitions | Can another analyst reproduce our counts? | Documented scope, statuses, filters, and counting units |
| Implementation effort | What work will our team need to maintain? | Integration tasks and unresolved questions |
MixRank’s job posting database
The MixRank job posting database contains 1.5B+ total postings since 2021 and 50+ fields. Job openings and hiring activity are connected to the hiring company’s record.
The database has 50M+ records updated in the last 30 days across all sources, with hourly update files. Delivery options include a REST API, bulk flat-file exports, and direct data-warehouse delivery.
You can read the API documentation or request an evaluation using your own employers, roles, locations, and historical requirements.
Frequently asked questions
What is the difference between a job posting database and job posting analysis?
A job posting database supplies structured records. Job posting analysis is the work of examining those records to answer questions about employers, roles, requirements, and hiring activity.
Is a job posting database the same as a job board?
A job board helps candidates find roles and apply. A job posting database supplies structured current and historical records for data products, research, and analysis.
What is a deduplicated job posting database?
A deduplicated database represents repeated advertisements so you don’t count them as separate opportunities. Check the provider’s record definitions and test repeated advertisements from your own market.
What is a historical job posting database?
A historical database keeps postings from earlier periods, including records that may have closed. That history lets researchers compare employer demand and job requirements over time.
Which delivery method should a buyer choose?
API access suits search, enrichment, and product workflows. Bulk files or warehouse delivery can suit large analytical workloads. Choose based on the system that will use the data, how often you need updates, and the engineering time you have.
A fixed test set and documented counting rules give you a practical basis for comparing job posting databases and reading the data they return.

Adewale Thomas
Adewale Thomas writes about technology and workforce data at MixRank, where he spends his time between sales-call transcripts, query logs, and the datasets themselves.



